Proceedings of The Third International Conference On Computer Science, Computer Engineering, and Education Technologies (CSCEET2016), Poland, September 2016

The Third International Conference on Computer
Science, Computer Engineering, and Education

Technologies (CSCEET2016)
September 19-21, 2016
Lodz University of Technology, Lodz, Poland
ISBN: 978-1-941968-38-32016 SDIWC
Published by: The Society of Digital Information

and Wireless Communications (SDIWC)
Wilmington, New Castle, DE 19801, USA www.sdiwc.net
Table of Contents
E- Learning
Virtual Labs and Educational Software as a Tool for more Effective Teaching
STEM Subjects
Identify Interface Design Patterns by Studying Intrinsic Designs
Education in Computational Science: Do Successful Examples Also Create
Success in Education?
Projects for Teaching Algorithmization in Primary Schools
Computer Engineering
Cluster Sampling for the Demand Side Management of Big Data
A Diagnosis and Prescription System to Automatically Diagnose Pests
Monitoring of events in SOA architecture for Real Time Financial Decision
System
Mobile Learning
A Study of Data Acquisition and Analysis for Driver's Behaviour and
Characteristics through Application of Smart Devices and Data Mining
Support and Improvement of the Educational Process in Regional Education in
Slovakia Through Tablet Classrooms
1
13
25
32
40
47
57
63
71
The Third International Conference on Computer Science, Computer Engineering, and Education Technologies (CSCEET2016), Poland 2016
Virtual Labs and Educational Software as a Tool for more Effective

Teaching STEM Subjects
Jn Gunaga, Rbert Janiga
Catholic University in Ruzomberok, Slovakia
e-mail: jan.guncaga@ku.sk, robert.janiga@ku.sk
Abstract
Significant development of information and
communication technologies and especially the
Internet boom brings new possibilities in
STEM education at all levels of the
educational process. We present in the first
part of the paper some results of international
ICILS study 2013 related to the digital
competences of pupils related to the computer
aided education, which use virtual reality and
computer simulations. The virtual reality is
distinguished by unique sorts of interaction,
that responds to users' behaviours and actions.
It is considered to be a new model of
computer-based learning that provides the
individual learner with a wider range of
scientific vision. In mathematics education it is
important to represent and visualize various
mathematical concepts, relationships and
practices,
respectively
algorithms.
Visualization is usually implemented through
different models. These models is possible to
present
through
virtual
and
remote
laboratories. These models may take the form
of real mathematical device, but may also be
implemented in a suitable virtual software
environment. Virtual modelling has particular
importance in the stages of cognitive processes
in STEM education.
Keywords: Virtual Laboratories, Motivation
in STEM subjects, simulations, e-learning,
ICILS.
1 INTRODUCTION
Great development of information and
communication technologies bring to the
educational process also the possibility of
ISBN: 978-1-941968-38-32016 SDIWC
using virtual laboratories especially in

teaching STEM (Science, Technology,
Engineering, Mathematics) subjects. For
their effective use by pupils and students
they need to have a certain level of
computer and information
literacy
(Computer and Information Literacy CIL). This problem has already been dealt
with many countries around the world on
different continents within the international
comparative research ICILS - International
Computer and Information Literacy Study.
This study compares the level of computer
and information literacy of different
countries. According to [9] Computer and
Information Literacy (CIL) is the ability to
use a computer to explore, create and
communicate in order to participate
effectively at home, at school, at work and
in society. CIL combines technical
competence and intellectual ability to
achieve communicative intent. CIL
construct consists of two components - the
collection and work with information and
the creation and exchange of information.
Each has some aspects or elements:
1. Gather and work with information:
Ability to use a computer and general
knowledge about it,
gathering information and their
evaluation,
work with information.
2. The creation and exchange of
information:
transformation of information
creation of information
sharing information.
According to [10] in international study
ICILS 2013 were involved 18 countries (Australia, Chile, Croatia, Czech Republic,
Denmark, Germany, Hong Kong (China),
South Korea, Lithuania, Netherlands,
Norway, Poland, Russian Federation,
Slovak Republic, Slovenia, Switzerland,
Thailand and Turkey. The research was
carried out in the eighth year of primary
school. This comparison showed that
Slovakia has significantly greater point
gains 517 points as an average 500 points
across studies ICILS. A similar level as
Slovakia has the countries - Germany,
Russian Federation, Croatia and Slovenia.
A significantly higher level than Slovakia
has - the Czech Republic, Australia,
Poland, Norway, South Korea. Lithuania
has an average of 494 points and three
countries - Chile, Thailand, Turkey has
significantly fewer points than 500 points,
what is the average of the study.
Interesting results for Slovakia provided a
study on the question on which school
subjects usually or almost every hour they
use a computer. 82 percent of the students
answered on informatics that was more
than the average of study ICILS. In
mathematics, it was 11 percent and
students in science subjects (Physics,
Chemistry, and Biology), it was 17
percent. These subjects belong to STEM
and is significant that these results are
worse for these items than the average of
study ICILS. It is therefore important to
conduct research and support activities to
increase the use of computers in STEM
subjects. For this aim we think that virtual
laboratories and educational software are
appropriate.
ISBN: 978-1-941968-38-32016 SDIWC
2 ACTIVITIES THAT SLOVAK

PUPILS USE COMPUTERS AND
INTERNET FOR, AT HOME AND AT
SCHOOL
According 70 % of pupils of SR involved
in study ICILS claim that they use for
computer at least 5 years, only 3 % of our
pupils use computer less than a year.
Approximately 95% use computers at
home at least once a week, at school 77 %
and at other places (e.g. library, internet
bar) 12 % of Slovak pupils.
Usage of computers and internet at school and
for school purposes
Tab 1. Slovak pupils use computers out-ofschool at least once a month for following
reasons concerning school:
52 %
Preparing papers and essays
51 %
Preparing presentations
41 %
Cooperation with other students of

the same school
35 %
Elaboration of worksheets or
exercises
30 %
Elaboration of tests
27 %
Organizing time and work
14 %
Cooperation with other students from

other schools
Tab 2. They use computer on most or every

class on subjects:
82 %
Informatics
18 %
Humanities (history, geography,

civics, etc.)
17 %
Natural science subjects (physics,

chemistry, biology)
16 %
Foreign language
Other subjects (practical and

technical subjects,
ethic/religious/physical education)
11 %
Mathematics
10 %
Art subjects (art / music / dance /

drama education)
10 %
Language (Slovak or Hungarian)
11 %
saving data and creating graphs

14 %
Using educational software, designed

to help with learning process
11 %
Creating computer programs, macros

and scripts
Tab 5. Slovak pupils use internet in out-ofschool time at least once a week for following
activities:
Tab 3. Slovak pupils learned following tasks at

school:
87 %
Typing messages using

communication software or social
network
60 %
Voice communication, on-line

chatting with friends and family
Decide where to seek for information

about given topic
53 %
Posting comments to on-line profiles

or blogs
71 %
Decide which information is

important to be included into the
project or task
47 %
Posting pictures, photos or videos to

profile on social network
42 %
Search for different types of
information about the given topic
Searching information for study or

school project
68 %
39 %
Mention links to the internet sources
Using WIKI or on-line encyclopedia

for study or school project
67 %
63 %
Sort the information gained from the

internet
29 %
Respond the questions from other

people on forums or web pages
59 %
Check the credibility of information

from the internet
24 %
Writing posts on own blog
21 %
Asking questions on forums or web

pages
14 %
Creating or editing web pages
84 %
Gain the access to the information

using computer
76 %
Present information using computer

to given audience or given purpose
71 %
Using computers and internet in out-of-school

time
Tab 4. Slovak pupils use computers in out-ofschool time at least once a week for following
activities:
Tab 6. Slovak pupils use computers for out-ofschool activities at least once a week for:
25 %
Creating and editing documents
22 %
Creating simple presentations
18 %
Creating multimedia presentations
18 %
Using graphic and drawing software
14 %
Using spreadsheets for calculations,
ISBN: 978-1-941968-38-32016 SDIWC
88 %
Listening to music
74 %
Watching downloaded videos or

streamed videos
69 %
Using the Internet to obtain news about

things they're interested in
61 %
Playing games
38 %
Reading reviews, comments about

things they would like to buy
35 %
Searching information about places

they want to visit or activities they
want to do
+ the percentage of students is significantly bigger

in SR compared with the average of ICILS
- the percentage of students is significantly smaller
in SR compared with the average of ICILS
* the percentage of students representation is
comparable in Slovakia and the average of ICILS
SIMULATIONS IN
MATHEMATICS EDUCATION
During the teaching process, the ICT can

support creation and processing of teaching
materials by ICT users tools, creation of
educational simulations and tests along
with an assessment of knowledge by
testing programs, and evidence and
evaluation of educational outcome.
An important aspect of ICT aided
education is the visualization. Within the
framework of mathematics education, it
can bring forward the following points:
An effective approach while looking
for the results, solving the problems
and discovering the very structure of
the model.
A visualisation of relations in the
model allows to infer new results in
other areas and fields of mathematics
and other subjects related to the
mathematics.
Support of mathematical competences
and basic competences in science and
technology; digital competences.
GeoGebra supports the realization of the
above aspects. This software connects
features of a computer algebra system, a
dynamic
geometric
software
and
ISBN: 978-1-941968-38-32016 SDIWC
spreadsheet. Its big advantage is a user

friendly nature and possibility to create
dynamic HTML websites with interactive
pictures (see [7]). Materials developed by
GeoGebra can be found on the website
GeoGebra Tube.
The latter one contains also some materials
created in Slovakia. We can find here
teaching materials for Mathematics at
school (http://www.geogebra.org/
en/wiki/index.php/Slovak).
These
materials, simulations and applets use the
interactivity and dynamical character of
software. Moreover, topics of materials
follow the Slovak curriculum from
calculus and geometry (e.g. topics of an
exponential and logarithmic function, cube
or hexagonal prism, constructions of
triangles, parallelograms). Naturally, the
website contains also materials for
teaching physics and computer science.
We show some examples now of them.
Figure 1: Concave mirror
Figure 2: Derivative of the function
4 VIRTUAL LABORATORIES AND

SIMULATIONS IN STEM
SUBJECTS
Practical experience is an important part of
the educational process. However, time
and economic resources often required to
establish and build science labs is out of
the scope for many institutions. The
solution to this problem can be found in
the use of virtual reality technology that
could allow the creation of a virtual
laboratory that will simulate the processes
and activities similar to those in real
laboratories.
Figure 3: Shutterstock/Macrovector ([16])
One of the unique capabilities of the

technology of Virtual Laboratory is
successful translation of abstract concepts
into visualized events with possibility of
user interaction.
One of the trends evolving in this direction
is an experimental learning or "learning
experience", where users are not just
passive
receivers
of
information.
Experimental learning requires a high
degree of interactivity
1. to share resources is becoming a reality,
improving the utilization of expensive
equipment,
2. to access to educational and research
material is made easier for students and
teachers,
ISBN: 978-1-941968-38-32016 SDIWC
3. to standard of scientific research are

established in areas where practical
experimentation is mandatory part of the
research,
4. to reduce travel costs and increase
productivity.
Virtual laboratories (VL) have as main
objective the creation of virtual learning
environments in order to provide further
practical teaching material and teaching
experience in disciplines such as computer
science. VL further provide communication and collaboration tools such as chat,
application sharing, etc. to simulate the
interaction of participants. Users access the
system using a web browser.
Well designed laboratory activities can
provide learning opportunities that help
students develop concepts. They also
provide important opportunities to help
students learn how to examine construct
scientific arguments and justify these
arguments in the community. In order to
achieve important but challenging goals,
the education system must provide time
and
opportunity
for
teachers
to
communicate with their students, as well as
sufficient time for students to perform
complex investigative tasks.
It is clear that there are serious
discrepancies
between
what
was
recommended for teaching lab classrooms
and what actually occurs in many classes.
Explanation may lie in the differences in
the perception of teachers and researchers.
For example, teachers it can be seen that
they do not have the time or skills
necessary to implement such methodology.
Educational research should produce
information to inform the development of
strategies, protocols and resources for
5
teaching and professional development for

teachers. Questions to be addressed include
how to effectively evaluate pupils' learning
how to engage students with different
abilities and knowledge to promote
effective learning environment laboratory.
Virtual
reality offers a pleasant
representation of information, interaction
with a system that does not require
advanced
knowledge
of
computer
technology and lower costs compared to
other technology.
At a time of rapid changes in science and

technology, teachers need to be informed
on current professional issues. The
development is very important and requires
attention not only teachers and their
professional associations, as well as
educational policy-makers at all levels of
schools, municipalities and governments.
6 USAGE OF THE VIRTUAL

LABORATORIES
5 BASIC FUNCTIONS
The main objective of the educational VL
is to provide all the simulation tools,
applications and conditions that will be an
effective space where experimentation,
communication and collaboration can be
used for maintenance and exchange of
knowledge. This means that the virtual
environment to be used for laboratory,
trying to simulate the learning process
from the very beginning until its end.
Users should simulate the real process as
realistically as possible.
Traditional laboratory was a part of an
integrated training comprises using a
laboratory
considerably
predefined
experiments that simulate the essential
phenomena taken from the real situation.
Unfortunately, this system has its flaws
and demanding requirements.
Virtual laboratories appear to be more
advantageous than the actual laboratories
particularly for specific experiments. There
are tests that can be simulated on
computers only.
ISBN: 978-1-941968-38-32016 SDIWC
1. In support of teaching theory

- Illustration / demonstration events
- Application of theory to real situations
- Demonstration of the theory of
constraints
- Interaction with phenomena in authentic
life situations
2. To create a set of knowledge
- On materials, equipment and technology
- The safety rules and practices
- Of special equipment and technology
3. To create a set of skills including
- Manual skills
- Critical observation, interpretation and
evaluation
- Diagnostic skills
- Planning and organization
- Solve practical problems
4. To develop attitudes that
- Stimulate interest in science
- Creating confidence in all areas
We assume that the VL can contribute to

the development of these skills as realistic
images of laboratory situations provide
opportunities for learning.
We define VL as a software simulation
experiments which data output is
indistinguishable from data from real
experiments.
Functions of measuring
equipment were clear
12
68
15
VL gave me the similar

experience as a real
laboratory
16
60
20
Instructions
to
measurements were easy
to understand
81
10
Figure 4: Virtual Labs in Fiber optics subjects

Tab 7. Questionnaire 1 after semester
VL opens up new challenges, such as

separation of imperfections in learning
technologies and methodologies and
developing strategies for the assessment,
providing reliable feedback on the students'
abilities. Feedback from students and
professionals brings many constructive
comments and suggestions. VL can also be
updated to include video guidelines for
adjusting. The video should include
instructions on how to connect the optical
circuit devices used in laboratories, and
instructions for measurement.
AD Absolutely Disagree, D- Disagree, A-Agree,

AA-Absolutely Agree
QUESTION
Yes
No
Were you able to identify the used

various connectors?
98
Were you able to perform the

experiment?
100
I was able to measure the output

peak power for different lengths of
fibers.
95
Laboratory helped me to better

understand the concepts.
100
I was able to set a different

wavelength ranges in Optical
Spectrum Analyzer.
100
VL has given me a similar

experience as a real experiment.
80
20
95
QUESTION
Yes
No
Control speed of laboratory was

sufficient.
100
Remote control software is simple to

use.
100
Virtual measuring device clear.
90
10
Visualization tool was clear.
90
10
Laboratory instructions were clear.
80
20
AD
AA
I was able to set the correct OSA and

gain characteristics of the spectrum.
15
77
Setting up wavelengths
was simple
33
59
My understanding of
fiber loss is now better
Measuring the strength

of optical signal in the
fiber was simple
Control of VL was
simple and intuitive
QUESTION
VL
increased
understanding of
LASER
my
the
ISBN: 978-1-941968-38-32016 SDIWC
19
66
72
71
34
26
10
It was easy to read data from the

devices.
100
VL provided the same experience as

a real laboratory.
80
20
Manual for the VL was easy to

understand.
100
The web interface was easy to follow

and understand.
85
15
7 VIRTUAL LABORATORIES IN
THE WORLD
Open University, founded in 1969, offers
distance studies and now has more than
240,000 students worldwide. Until the late
90's, the science courses provide students
with kits that include microscopes, printed
circuit boards, chemistry sets, containers or
even lasers. Students experimented at
home and then sent the unit back. But it
was expensive and not very practical.
Currently, almost all laboratory work is
available online via the university Open
Science Laboratory. Like many working
scientists, students can collect real data
from remote controlled devices - for
example, - spectrometry for the
identification of elements and isotopes, and
00:43-m telescope in Majorca in Spain.
Students can also explore realistic
simulated data from tools like a virtual
microscope with which we look at the high
resolution images instead of real samples.
Paulo Blikstein, director of Transformative
Technology Lab at Stanford University in
California, goes on with a new generation
of digital laboratory courses. One example
of using remotely controlled centralized
tools for biological laboratory - a project
developed with Ingmar Riedel-Kruse is
bioengineering at Stanford. "The idea is to
have an area of 10,000 petri dishes, each a
ISBN: 978-1-941968-38-32016 SDIWC
few millimeters wide and a robot that acts

like a inkjet printer," says Blikstein.
"Students should operate the robot," Go
into my bowl and add drops of X, "and the
camera would see what happens."
One of the key elements in effective
mentoring is a conversation, says Graesser.
"When people only read books or listen to
a lecture receive a smattering," Graesser
and his colleagues, for example, developed
a system based on "trilogies" in which the
student interacts with two computer
animated agents - teachers or students.
Real students can interact with the teacher
directly, or they can deepen their
knowledge by learning with peers.
One of the biggest barriers to wide
adoption of systems for the teaching of
practical skills is that many single
experiments are not associated with
MOOCs or anything else. Once the project
is completed, "you can put your
application on any website where they can
be almost none found," said Daphne
Koller, co-founder of the largest MOOC
company courser in Mountain View,
California.
Courser is trying to change, says Koller not only by encouraging experimentation,
but also rewriting their own software in
order to offer practical applications next
lecture courses. It is hoped that creating a
single market for these applications, and
gives them a much wider exposure.
We are at the point where we can start to
think about learning a whole new way.
Computer science has invaded the
educational process and is providing us
with many opportunities to exploit. An
additional challenge faced by STEM
educators has been the integration of
Inquiry Based Science Learning (IBSL) in
teaching. While the use ICT has already

penetrated in Inquiry Based Teaching, we
have yet to create and implement STEM
lesson plans that promote the development
of methodological skills and competencies,
investigation through experimentation,
teamwork and communication among
students through collaborative activities.
In recent years, Inquiry-Based Science
Education (IBSE) has proved its efficacy
in education by expanding on traditional
lessons and motivating students to actively
participate in science [11]. IBSE methods
and digital technologies support necessary
educational innovations and can be the
catalyst for change in educational patterns
(in regard to its form, space, functions,
services, tools, roles, procedures) [12].
Virtual laboratories are an essential digital
tool. In fact, many European schools are
equipped with computer classes, tablets
and high-speed internet connection while
using a huge variety of web-based learning
applications,
simulations
and
visualizations. [3]
8 DESCRIPTION OF VIRTUAL
LABORATORIES
ENVIRONMENTS
Virtual laboratory environments can be
divided into following categories [16]:
popular in science subjects. They are small

in size and easily transported and they can
be used regardless of the operating system
type.
Virtual labs
Virtual labs (virtual laboratories) simulate
a virtual operating system, the computer
screen, Science laboratories, exploiting the
potential offered by modern media
technology
key
feature
technical
interaction and direct and plausible
manipulation of objects and parameters.
Virtual Reality Laboratories (VRL)
VRL workshops are computer based and
highly interactive. The user becomes a
participant in a virtually real world, in an
artificial
three-dimensional
optical
environment. These workshops are
essentially an interface high level including
real time three-dimensional simulations
through different sensory channels.
Laboratories Controlled
(Remote Labs)
by
Distance
Workshops controlled remotely (remote

labs, otherwise known as online labs or
workbenches) include real experiments
conducted from a distance with the use of
telecommunications, while the user uses
this technology from another location.
Simulations
Simulations are imitations of operating
systems through time, via computers.
These represent a process on the basis of a
model that is cheaper, faster, less risky and
more affordable than the real process.
Network applets
Most of the virtual laboratory software

consists on computing applications running
on the local users computer, for speed and
security reasons. They can be operated
remotely. An example are those based on
applets or robotic workshops (remote labs)
that can accept commands via the Internet
[12]
The applets are experimental devices in

small virtual laboratories and are quite
ISBN: 978-1-941968-38-32016 SDIWC
9 BENEFITS OF USING VIRTUAL

LABS IN TEACHING AND
LEARNING SCIENCE
Virtual labs can be very useful in the
teaching of Science, particularly in cases
where:
The experimental activities are to be done
quickly and do not easily allow
observation and safe measurement,
-
the experimental process is very

slow and / or complex and not
compatible with the teaching time
available,
the experiments involve risks to the

health and physical integrity of
learners and/or
the learning
modeling.
activities
require
Virtual labs support IBSL in learning

science:
Laws in science arise from a detailed
observation processes, with clearly more
chances of clarification, understanding and
acceptance if regarded in detail.
It
encourages
collaboration
and
communication between teachers and
students. STEM teachers participate
actively in the learning process: asking
questions, trying to find answers,
organizing procedures and commenting on
them, helping in formulating conclusions,
understanding
their
mistakes
and
highlighting any misconceptions.
But what are the differences between real
life experiences and those formed by
representations in a computer screen?
ISBN: 978-1-941968-38-32016 SDIWC
With VL, students acquire a tool with

which to experiment without limitations of
space or time. They are available all year,
as opposed to school laboratories, limited
to a specific place and for a limited time.
[4].
The use of virtual environments makes
students acquire better computer skills,
which can be considered skills for lifelong
learning. The use of these technologies
also bring together different STEM
subjects and provides with great resources
for more inclusive workshops [8].
10 CONCLUSIONS
Virtual Laboratories (VL) is environment
for students experiments and it is one of
kind of using dynamical character of the
educational software in the STEM
subjects.
Different methods of the
visualisation and simulation we can use in
different STEM subjects. How we
described above, the virtual Lab Concept
was defined as "laboratory experiment
without real laboratory with its walls and
doors. It enables the learner to link
between the theoretical aspect and the
practical one, without papers and pens. It is
electronically programmed in computer in
order to simulate the real experiments
inside the real laboratories." (compare
with [13] and [14]). For other applications
in the aviatics field see [1].
There exists many simulations in the
mathematics education developed by
GeoGebra. The GeoGebra research
community has already international
character which brings the opportunity to
exchange the experiences in the field of
motivation of pupils and students in
mathematics and STEM education.
10
Visualisation through simulation in STEM

education helps students to understand new
notions, relationships between them and
incorporate this new knowledge in existing
structure. It is important for development
of different levels of thinking in science
education. Simulation is possible to use in
many e-learning courses for STEM
education and other fields such aviatics
(see [2] and [7])
Real experiment is impossible to change
through virtual experiment in educational
process. The role of the virtual experiment
is such supporting tool for better
understanding
of
principles
of
demonstrated phenomena. If teacher has
possibility to realize virtual experiment
using computer simulation, he can develop
important students skills needed for their
professional life not only in STEM field,
but also in other fields (see also [15]).
[3] Dikke D., Tsourlidaki E., Zervas P., Cao Y.,

Faltin N., Sotiriou S., Sampson D. Golabz:
Towards a federation of online labs for inquiry
based science education at School.
[4] Doukeli M. (2012). Virtual labs in teaching
physics in secondary school. Research paper for
Master Degree. University of Piraeus at department
of Digital Systems.
[5] Fischer, J., Mitchell, R. & del Alamo, J.
(2007). Inquiry
learning
with
WebLab: Undergraduate
Attitudes
and
Experiences. Journal of Science Education and
Technology,16 (4), 337-348.
[6] Harms, U. (2000). Virtual and remote labs in
physics education. Proceedings of the Second
European Conference on Physics Teaching in
Engineering Education, Budapest, Romania (pp. 16).
Acknowledgment
[7] Hohenwarter, M., & Lavicza, Z. Gaining

momentum: GeoGebra inspires educators and
students around the world. GeoGebra The New
Language for the Third Millennium. Zigotto
Printing & Publishing House Galati-Romania,
Vol.1 No.1. 2010, p. 1-6.
Supported by grants KEGA 002UJS-4/2014,

KEGA 017KU-4/2014 and TA 040 313 76 Vzkum/vvoj metodiky vcviku leteckch
specialist L410 UVP-E20
[8] Majherov, J., Palsthy, H., Janigov, E. Pupils

in the Virtual World and Education, 2014. Lecture
Notes in Computer Science. - ISSN 0302-9743,
Volume 8730, (2014), p. 112-123.
References
[9] Predstavenie tdie ICILS. Available from

electronics address: http://www.nucem.sk/
documents//27/medzinarodne_merania/icils/Predsta
venie_tdie_ICILS.pdf
[1] ernk I., Petrukov H., Majherov J.

Didaktick aspekty vcviku leteckch pecialistov,
PERNERS CONTACS, Elektronick odborn
asopis
o technolgii,
technice
a logistice
v doprav, Dopravn fakulta Jana Pernera,
Univerzita Pardubice, Katedra technologie a zen
dopravy, Didaktick aspekty vcviku leteckch
pecialistov, p. 22-31,aprl 2015, ISSN 1801-674X
[2] ernk I., Krlk V.: E-learningov podpora
vuby programovania In: Inovan proces v elearningu : recenzovan zbornk prspevkov zo 7.
vedeckej medzinrodnej konferencie: Bratislava 26.
marca 2013 / eds. Martin Blahuiak, Janette
Brixov, G. Kristov, Miroslav Krjak, Eva
Rakovsk. Bratislava: Ekonm, 2013. - ISBN 97880-225-3610-3, CD-ROM, p. 1-4.
ISBN: 978-1-941968-38-32016 SDIWC
[10] Krtka sprva ICILS 2013 Slovensko.

Available
from
electronics
address:
http://www.nucem.sk/documents//27/medzinarodne
_merania/icils/publikacie/ine/Krtka_sprva.pdf
[11] Niederrer et al. (2003). Research about the use
of information technology in Science Education.
Education research in knowledge based Society.
Kluwer Academic Puplishers.
[12] Rocard, M. (2007). Science education NOW: a
renewed
pedagogy
for
the
future
of
Europe. Luxembourg:
Office
for
Official
Publications for the European Commission.
Available at: http://ec.europa.eu/research/science-
11
society/document_library/pdf_06/report-rocard-onscience-education_en.pdf.
[13]
Sampson,
D.
(2010).
Instructional
Design. Course Lectures. University Piraeus 2010.
[14] Tselfes, B. (2002). Trial and error: The
workshop
on
the
teaching
of
Science. Athens:Island. Education and Training
Sector (TEK). Training material for teacher training
Issue 5: Sector PE04. CTI. Available from
electronics
address:
http://axis.teikav.edu.gr/
pake/Enotita_7_Logismika_PE04/AMAP_Anoikto
_Mathisiako_Perivallon / AMAP-Intro.pdf .
[15] http://onlinelabs.in/
[16]http://blog.scientix.eu/2015/08/20/virtuallaboratories-in-teaching-and-learning-science/
ISBN: 978-1-941968-38-32016 SDIWC
12
Identifying Interface Design Patterns by Studying Intrinsic Designs

Ken D. Nguyen and Muhammad A. Rahman
Clayton State University
2000 Clayton State Boulevard, Morrow, GA 30260
KenNguyen@clayton.edu and MuhammadRahman@clayton.edu
ABSTRACT
Designing an effective user interface for a new
system is challenging even for the most experienced
designer. In this study, we present a strategy to find
a user interface design pattern for various problem
domains through a systematic process. Our strategy
arises from studying and analyzing several popular
human-centered interface designs in their entirety of
evolution. We apply our technique to redesign
Walmarts online-store checkout web-forms, reduce
their complexity by almost 80 percent, and greatly
improve its performance without compromising its
functionality.
KEYWORDS
Interface design, user interface, design pattern,
digital interface.
1 INTRODUCTION
User Interface (UI) design has traditionally
been the most important factor for designing
applications from analog systems to digital
systems. An interface of a system is the primary
means for the user to interact with the system;
and it is the gateway where engineers can hide
the details and complexity of the system
infrastructure and still provide the users with
most, or all, functionalities the system is
capable of. The user interface of a system is a
great way to allow the user to abstractly see the
system as a much simpler model.
In reality, the interface designers give users a
wide range of experiences from intuitive and
productive to cumbersome and wasteful.
However, to design a successful user interface
ISBN: 978-1-941968-38-32016 SDIWC
defined as a simple, intuitive, and to effectively

allow the users to complete their tasks with
minimal interactions - is a great challenge.
Show everything at once, and the result is
chaos. Dont show everything, and then stuff
gets lost [1]. And the best interface is no
interface [2], i.e. the system is ubiquitous and
capable of recognizing the users intentions
with precision. Most systems are designed to
solve common problems, and almost always
there is an interface for a similar system in the
same problem domain. There are very few new
or emerging systems that need a completely
new user interface design. In this study, we
analyze many popular interface designs across
domains to recognize patterns for successful
user interface designs. We also illustrate our
methodology to design the next generation
interfaces for the smart thermostat NEST, chipequipped credit cards, and Walmarts online
checkout forms, the worlds largest retailer.
2 QUALITATIVE MEASUREMENT
The popularity of a user interface of a system or
product is largely dependent on how the system
or product is introduced to the world. That
involves the marketing strategy, the brand
recognition, the organization fan-base, etc.
Thus, the popularity or success of a product
may or may not implicate that there is a direct
relationship between the user experience with
the product interface and the success of the
product. For example, IBM [3] in 1992
introduced a touch-screen smartphone named
Simon with very limited success (50,000 unit
sold), and there are many other companies such
as Palm, Qualcomm, Ericson, Nokia etc. that
13
faced similar outcomes; however, Apple, Inc.

introduced iPhone with a multi-touch interface
in 2007 with an ongoing success, which is
largely contributed to iPhones intuitive user
interface [4].
Traditionally, interface designs should be tested
vigorously starting with informal testing and
evaluation tests to recognize, prioritize, and
elaborate the users wants to needs. All
interface designs need expert-reviews through
heuristic evaluations that focus on (1)
established heuristics such as Nielsens 10
usability heuristics[4], or Shneidermans 8
golden rules[5] or Tognazziniss first
principles, etc., (2) guideline review that checks
a design against established guideline
documents, (3) consistency inspection that
focuses on consistency at various levels from
the component level through the system level,
(4) cognitive walkthrough that focuses on userand task-oriented scenarios where experts go
through typical user tasks, (5) and formal
usability tests and inspections where the experts
evaluate and justify piecemeal designs through
various user tests to identify fatal design flaws.
There are so many well-known strategies and
methodologies for designing and developing an
effective user interface, especially those
modern methodologies that employ context into
their designs as seen in [6] and [7], however,
there are so many existing systems with not so
great interface designs built by great designers
of many world leading organizations and
businesses.
Thus, in this study, we define a successful
interface design as an interface design that is
intuitive and directly allows the user to
complete their tasks effectively with minimal
learning curve and operational steps. Our
measurement is based on the actual tasks to be
completed, the classical solution, the minimal
operational steps, and the actual interface.
3 CASE STUDY
ISBN: 978-1-941968-38-32016 SDIWC
Since our goal is to identify a successful user

interface design, our approach will be in the
opposite direction of the traditional approach.
We examine existing designs and trace along
their evolution path to identify which
functionalities and features have survived
and/or reoccurred in popular existing designs.
What we found may not be a design pattern for
a specific problem but rather a pattern to design
a user interface that will likely be accepted by
its users in the domain of that problem. We
have analyzed dozens of systems; however, we
limit our scope to only 6 groups of user
interface designs for 6 domains of popular
problems.
3.1 Telephone Systems
Direct voice communication has been a part of
human nature, even before the invention and
patent of the telephone in 1876 by Alexander
Graham Bell [8]. The most intrinsic
communicating method has been exchanging
information vocally. It has been done
continuingly for thousands of years. We can
clearly observe this technique in a crowd
gathering in a busy shopping mall, outdoor
concert, or sport event, where two or more
people who know each other may call a name
out-loud to get the attention of each other and
then start talking softer when the involving
parties are aware of each other. Strangers may
communicate with each other in exactly the
same way; where a person may call on another
and use surrounding objects to get the others
attention such as the lady in pink standing on
the left of the podium next to the clown carrying
red balloons or scream Hey Joe and
jumping up and down waving their hands to get
the other party's attention. In a more formal
environment as in a train, bus, or in a
conference, one may ask the nearby person to
get the attention of another such as I want to
talk to the gentlemen in white suit, can you get
him for me please, and the message will be
relayed to other persons closer to the other
party until it reaches the intended individual.
14
The fundamental idea is that one party knows

who he/she wants to communicate with and
he/she relies on the voice, personal information
and/or surrounding context to initiate the
communication. Using ones vocal and hearing
capability is intrinsic to human. The original
design of the telephone system closely follows
this intrinsic communication technique where
the telephone has a hearing cup and
microphone - the caller will tell an operator
who he/she wants to talk to and the operator
will establish a link between the two parties.
This design is purely simple and intuitive,
however, it is infeasible for a system with many
users. Phone numbers are deployed for a larger
system, with many users, to map each
individual to a number. Humans use context to
associate a person with an event and recognize
relationships between subjects so they can
quickly recall the reference. All of these
contexts have become an abstract number - a
quick fix introduced by engineers because of
the incapability to decode humans contexts.
The users trade these contexts for a number to
remember, a keypad to dial, and a contact list to
keep.
In modern smartphones, the personal contact
list is stored in the phone and can be easily
retrieved. Most smartphones now have the
voice recognition engine that allow the user to
give vocal command such as call my wife
and the phone will search the personal contact
list and make a call accordingly. Thus, the most
advanced solution resembles the first
mechanical solution which closely follows the
human intrinsic communication model. Figure
1 shows a representation of these phones.
Figure 1. A wooden wall telephone with hand-crank

[left], an Internet Protocol (IP) phone [center], and a
smart phone (iPhone)[right]
ISBN: 978-1-941968-38-32016 SDIWC
The main pattern of voice communication is:

vocally initialize the conversation, vocally
communicate, and vocally terminate the
conversation.
3.2 Typing Keyboards
Long before the invention of keyboards,
humans have been scripting onto cave walls to
describe events. And in the modern time, we
are doing the same thing with slightly different
styles and materials. The first typing device was
designed and patented in the 1700s, and the
first manufactured typing devices came around
1870s.
Later on, the invention of the keyboard allows
the users to mass manufacture their messages
easily. In the early days, typesetting, scripting,
and typing were done through an operator,
where the user/author provided the operator a
script to make imprints. Eventually, typewriters
and computers become popular and accessible
for personal and individual use.
Figure 2. 1852 John Jones Mechanical Typographer

[left] and 1870 Remington foot-pedal sewing machine
type writer [center], 1961 Electric I Typewriter by IBM
[right]
Currently, the modern digital computers are

dominating most of the printing revenue; the
keyboards look very much the same whether it
is a physical keyboard or a virtual keyboard
displaying on a devices screen.
While keyboards are different both in physical
presentation and mechanism to make imprints,
their functionality is identical allowing the
user to script consistently. The current mobile
technology and touchscreen has a new realm of
15
keyboards. Figures 3 shows different types of

modern keyboards.
Figure 3. One-handed keyboard [left], computer [center

left], Virtual keyboard on a smartphone [center right] and
hologram projected keyboard [right]
Nowadays, users can dictate their messages on

mobile devices that are equipped with virtual
keyboards and integrated voice commands. As
it has been shown, the fundamental operation
need for the keyboard is for humans to encode
their message either through dictation or by
directly choosing the symbols.
The main pattern of keyboarding is: persistent
transcription by selecting familiar characters/
words/ sentences/ paragraphs, preferably in
bulk or premade template.
3.3 Thermostats
The thermostat is another popular and
interesting interface. The main functionality of
the thermostat is to trigger a signal when the
temperature is reaching a preset point. The
main use of the thermostat is to activate the airconditioner or the heater to keep the
temperature in a preferred range. The
traditional mechanism to control temperature,
especially in a closed building, is to open the
place and let the air get through when the place
gets hot or close the space and provide more
heat by burning some wood/fuel when the place
gets cold. Humans in the modern day are doing
the same thing.
The first invention of the thermostat is credited
to Cornelis Drebbel who designed a regulator, a
column of mercury and a system of floats and
levers to keep the temperature of a portable
ISBN: 978-1-941968-38-32016 SDIWC
oven steady. The first patented thermostat was

designed by Albert Butz in 1885, patterned in
1886 as a thermo-electric regulator.
Thermostats come in many shapes, sizes and
functionalities. The newer thermostats are built
on digital circuits, which are capable of
accommodating different users settings for
preferred temperatures at different times as seen
in Figure 4. While the addition of digital
computer processing power into the thermostat
is a great idea, it did not work out very well due
to the cumbersome programming task. And the
programmable thermostats do not save energy
[9].
Figure 4. Thermo-electric regulator[left], 1968 model

w884 with dials to control both temperature and humidity
[center left], and 1986 - chronopher III - programmable
thermostat [center right], NEST- an intelligent thermostat
[right]
Lately, the introduction of NEST, as seen in

Figure 4 (right most), an intelligent thermostat
that allows the users to set the temperature
exactly the same way as they would do on the
1968 thermostat model w884. The difference is
the NEST uses its computing capability to
record the users preference and auto program
itself [NEST has many other advanced features
that we are not addressing in this study]. This
feature certainly could have been implemented
over half a century ago on the first model of the
programmable thermostat, however, the simple
concept was overlooked leading to hundreds of
unwanted complicated interfaces.
The main pattern of operating the thermostat
is: simple adjustments of temperatures as
needed. Precision is not important, but similar
setup is expected daily or weekly.
3.4 Credit Card Payment Methods
16
Since the very early days, humans have two

systems of trading - direct trade and credit
trade. In direct trading, an item is traded for a
service or another item with equivalent value.
In credit trading, one party may not have a
physical item or money for trading but he/she
has reputation or proven that he/she has the
capability and willingness to repay with similar
value of the item that he/she wants to get.
Credit trading is flexible and convenient; thus, a
third party is often needed to help authenticate
and manage transactions. Credit cards,
represented as charge plates and credit coins,
had been used since the 1800s until the plastic
card was introduce in the 1960s [10].
Advancements
in
telecommunication
technology have a great influence on how credit
trading transaction is done, which leads to
different development of interface for
debit/credit card authorization. The plastic
cards are now equipped with a magnetic strip
containing the owner information that can be
extracted easily by a magnetic card reader [as
seen in Figure 5], and this information can be
transmitted
through
the
existing
telecommunication system for processing.
Modern card readers are usually integrated with
the vendors point of sale computer system to
communicate with the card issuer for
authentication and approval of each transaction.
The interface for processing credit/debit cards

usually involves the authorization step: The
cardholders pay for their merchandise by
presenting the card to the merchant for
processing. The merchant then sends the credit
card information along with the purchase
details to the card issuer for verification and
approval. Common information needed for a
transaction authorization are: card number,
expiration date, amount request, and possibly a
pass-code on the card that is not stored in the
magnetic strip (for added security protection).
Recently, card readers can utilize the Internet

and the cellular telephone network for
transmitting their data. Thus, leading to
development of card readers that can be
attached to a mobile device. Card issuers also
equip the new cards with a micro- chip to
prevent vendors from storing the card owner
information electronically. Ironically, the user
information is still imprinted on the card.
The most popular interface for card processing

is at the point of sale where the merchants
obtain the card information for authorization
and further processing. Traditionally, the
merchant authenticates a card by calling the
card issuer with the card information and have
the cardholder verbally confirm their identity
and the transaction with the card issuer. This
technique is the best method against theft, and
it is the method of choice for large transactions
today.
Figure 5. Different type of credit card readers
ISBN: 978-1-941968-38-32016 SDIWC
The Internet also leads to the development of

online stores that prompt the card holders to
enter the card information. Despite the
differences in the platforms used, all the
information needed for a card authentication is
very much the same.
The most notorious development of the mobile
system would be ApplePay and AndroidPay,
which allow the users to enter their credit card
and banking information into a capable phone
prior to the first purchase. The user makes a
payment by bringing his/her mobile device next
to the reader and hitting a button. The token is
transmitted wirelessly to the reader, which in
turn sends the token along with the transaction
details to the server for authentication. Neither
the users personal information nor the funding
information is stored on the mobile device or
transmitted to the reader, thus securing the
transaction even if the token is intercepted by
17
an unwanted party. ApplePay uses a sensor to

check owners fingerprint when he/she hits the
button to make a payment, while AndroidPay
currently uses a password for older devices and
fingerprint sensors on new devices.
Since most people tend to use their mobile
phones several times daily, their phones are
more readily accessible and get more attention
than a credit card in a wallet. In addition,
ApplePay and AndroidPay will give an
immediate notification of any transaction to the
phones alerting the users of a new transaction
and making AppplePay and AndroidPay more
secure.
Figure 6. ApplePay [left] and Androidpay [right]
The main pattern of credit payment methods is:

simple and secured mechanism for exchanging
wealth. Validation of transaction authentication
with actual authorized user of each transaction
is expected.
3.5 CAPTCHA: Human-Recognition
The capability to distinguish humans actions
from others is instrumental in many situations.
For example, to know if someone is at the other
side of a closed door, one could ask who is
there?. And if there is a human voice
answered, we are sure that there is someone at
the door, and most of the time we would know
who that is if we can recognize a familiar voice.
In many situations, a person can recognize a
loved one from reading a few lines of text on a
piece a paper hidden in a bread roll smuggled
into a highly secure prison or from a washedashore piece of paper in a bottle. Humans can
do this, with little or no effort, because they can
recognize the meaning of the action and the
context in which the action occurs, either in the
ISBN: 978-1-941968-38-32016 SDIWC
past or in the future. This capability, while

intrinsic to humans, is a very difficult process
to automate. And it is one of the critical tasks
on many online systems to avoid systematically
abuse by computer programs. For example, an
email registration system would give each user
a chance to input his/her data to obtain an email
address. A computer program can easily submit
a large amount of random data to the system in
a short amount of time to register for many
email addresses. It is an impossible task for
humans to sort through, recognize, and
eliminate the illegitimate submitted data. It is
very difficult to implement a program to
distinguish a human action from a computer
generated action designed to simulate a human
action.
One popular mechanism to detect human action
from a computers is through a CAPTCHA,
which is short for Completely Automated
Public Turing Test to Tell Computers and
Humans Apart and introduced by Luis Von
Ahn et. Al. in 2000[11]. Through a CAPTCHA
system, the user is presented with a task to
resolve, and the result is evaluated for
correctness. Some popular CAPTCHAs are:
solving a simple math problem, script/type
down a message from an audio/video segment,
recognize some distorted text, arrange
humorous cartoon pictures in the appropriate
sequence, etc.[12] and [13]. These CAPTCHA
systems are easily programed to generate
CAPTCHA tests and evaluate the results of the
tests.
There are two major problems of existing
CAPTCHA: (1) it is difficult and time
consuming for humans to do the task right and
(2) the computer programs can solve most of
these tasks effectively. Thus, the existing
CAPTCHA becomes an extra burden for the
human and an ineffective deterrent tool that
may be able to discourage novices from
abusing the system. Current advancement in
artificial intelligence enable computer programs
18
to crack up to 90% of the CAPTCHAs used in

Google, Yahoo, and PayPal websites [13].
The latest development of CAPTCHA is No
CAPTCHA from Google, where it prompts the
user to click on a checkbox confirming that
they are not a bot - a computer program.
Obviously, clicking on a checkbox is easy,
which requires little effort both from the human
and a computer program. Google tracks the
movement of the mouse and compares that to
human patterns of mouse movement to detect
similarity and compare that to other knowledge
about that particular user Google has collected.
If Google does not obtain a high enough
confident level that the action is being taken by
a human, it will prompt the user to click on
images with the same context with a given
sample image, from a given set of images. For
example, the user is asked to pick out turkey
images from a set of Thanksgiving images
including turkey, bread rolls, jam, etc. given in
Figure 7. The pictures taken of different types
of turkeys in different angles are intuitive for
human to recognize with little effort but are
time consuming for computer programs to
detect. Googles No CAPTCHA system, while
still is very primitive comparing to humans, is
the closest thing to how humans would
recognize a humans action - ask for a simple
task and recognize the human action by the rich
knowledge common and shared among humans.
Most of human operations are done with an

intention to accomplish a task, such as pressdown on the gas pedal to accelerate a moving
vehicle, dropping a letter to the mailbox to mail
a letter, shred a sheet of an important document
to destroy it, or turn/off the TV, etc. Humans
operate with the same intention when they are
on computers. When a user selects a file, then
presses the Delete key on the keyboard or
right-clicks on the computer mouse and chooses
Delete, his/her intention to remove that file is
clear; thus, the computer should perform the
requested action. Most Unix/Linux based
systems operate this way. Unfortunately, the
vast user-friendly computer systems and
applications that rely on Graphical User
Interface (GUI) such as Windows are generally
not designed to work that way. This case study
examines the alert and confirm sequence,
however, we should be able to relate to all other
computer systems in a similar way.
Working on a computer system sometimes is a
cumbersome task when the users intention is
clear, and the system keeps asking for
confirmations. For example, on Windows
system, when a user wants to organize and
clean up his/her desktop space, she/he deletes
some of the un-wanted files. The system will
prompt every time a file is deleted as seen in
the Figure 8 [left].
Figure 8. Alert and confirmation of deleting file in

Windows
Figure 7. Googles no CAPTCHA reCAPTCHA

The main pattern of CAPTHCA is: simple and reliable
way to verify a human is actually initiating the request.
When the user is removing a large file, another

confirmation will popup alerting the user that
the file is too large to fit in the Recycle Bin as
in the Figure 8 [right]. Finally, when the user
tries to empty the Recycle Bin, another alert
3.6 Alerts and Confirmation Sequences
ISBN: 978-1-941968-38-32016 SDIWC
19
pops up, telling the user that the files will be

deleted permanently.
application. Thus, this alert system is more

helpful and natural to the way humans operate.
While these alerts have a good intention to let

the user know the consequence of his/her
action; however, it is annoying and
unproductive, when the users intention is clear
and the computer stops the user from doing
anything else unless he/she responds to the
alert. The only usefulness of this alerting
system is that when the user accidently
performs an unintended operation such as when
his/her palm touches the touchpad causing a file
to be moved into a nearby directory
unfortunately, the current Windows system
would not give any indication for this type of
operation. Thus, the accidental operations are
slipped through without notice and the intended
operations are delayed at every step with these
mindless alerts. This model of alert, confirm,
and re-affirm operations exist in all computer
systems, varying by some pre-determined
perception of how important an operation is or
how severe the consequence of the operation is.
These alert and confirmation sequences are also
popular in most web-based applications, in
which the developers put them in as protection
mechanisms regardless of whether it is actually
needed or not.
The main pattern of alert and confirmation is:

notifying the user immediately of critical
situations without interfering with the users
current interaction with the system. The user
can take action or ignore the notification.
A more accepted alert system is the one that is

discrete and does not interfere with the users
task. For example, Googles applications show
a small non-interfering alert message
immediately after an action is performed
allowing the user to revoke the operation if
needed. The message will go away in a few
seconds and does not interfere with users
current task.
In this approach, the users action is always
performed, and the user is given a notification
of the action and given a chance to revoke the
operation if the user action is done by mistake.
The notification does not need any users
interaction and does not hinder the use of the
ISBN: 978-1-941968-38-32016 SDIWC
4 INTERFACE DESIGNS WITH NEW

METHODOLOGY AND EVALUATION
To test our methodology, we have two
independent groups, each with 15 members,
evaluating Walmart.com (the worlds largest
company by revenue according to the Fortune
Global 500 list in 2016 [14]) retailer in online
order interfaces, the NEST thermostat and the
new chip-equipped credit cards in two rounds.
The first round is a blind test for user
experience. The second round is reevaluating
their user experience infused with our design
pattern discovery and then design the next
generation interface. The first group consists of
individuals who have some direct or indirect
experience with the test subjects (shopping
online, the NEST and the chip-equipped credit
cards).
4.1 Designing an effective online order webforms
Web-forms for online order checkout is
fundamental to all online transactions. We have
amateurs who only get a few minutes of
training to improve Walmarts checkout webforms (Walmart.com). Walmart, Inc. is the
largest retailer in the world which employs a
very large software design and support groups.
Round 1: In this round, each member is asked
to timely make an online purchase of an item
showing on the front page of Walmart.com as a
new patron. Each member records the number
of steps he/she has to go through and the data
must be provided at the level of mouse clicks
and keyboard strokes, excluding any mistakes.
20
All members are able to proceed to the page

where they can submit their order after going
through the following interfaces: (1) provide
their name, email and password (twice) to
create a profile, (2) choosing a shipping option,
(3) provide a complete name (again) and
shipping address, (4) provide full detailed credit
card information and billing address, (5) review
the order and submit. A screen capture of one
of these interfaces is shown in Figure 9.
purchase from Walmart.com, and they all

express that it is not as good as they thought.
For this round, the members are asked to design
a new online processing interface for
Walmart.com to minimize the users
interactions and still achieve two key-tasks: (a)
complete a purchase and get a customer profile
for the seller. Majority of the new interface is a
single step simple web-form that prompts the
users to provide (1) a credit card number and
the corresponding postal zip-code. The postal
zip-code is used to authenticate the card and
obtain the customer residing city and state, (2)
shipping street and the building number the
users can change this information if they want
to send to a different address, and (3) an
optional field for email or phone number if the
user wishes to get a receipt copy forwarded to
the given email or phone (as a text message).
The new interface is shown in Figure 10.
Figure 9. A checkout web-form from Walmart.com

taken on August 25, 2016
When asked about the purchase experience, all

members express that their checkout
experiences are fine besides some information
are asked twice such as name and address.
When they are asked to redesign the check-out
web-form for Walmart.com, the new interface
is similar to the existing one where the
redundant fields asking for name and address
are removed.
Round 2: All members are asked to describe a
scenario where they make a quick purchase of
an item (not online). Most members describe
their experiences of a purchase at a gas station
or a convenient store where they pick up an
item, swipe the credit card, and then sign the
pad; or pump some gas where they insert a
credit card, put in a pin or their postal zip-code
to complete the transaction. The members are
asked again for their experience with the online
ISBN: 978-1-941968-38-32016 SDIWC
Figure 10. One-step checkout web-form redesigned for

Walmart.com. A change in the auto filled shipping zip
code field indicates the billing address is different than
the shipping address.
When the user submits an order, the transaction

is completed. The user is presented with the
complete order details where he/she can make
changes if needed. At the server, the order will
be delayed 30 minutes after the last change
made to the order before sending to the
processing center which is also a common
practice in the industry. With this simple
interface, the members justify the users will
have minimal interaction similar to their
transaction at a gas station with the addition of
providing a street address for shipping. And the
21
company still be able to create a profile for the

customer by sending the customer a computer
generated link/password to access their online
orders on the electronic receipt to the email
address or the phone number the user has
provided. The user is not asked to enter his/her
name, and the justification is that the name can
be obtained by looking up the phone number,
the cashier can obtain the name at the time of
pick up in the store, or the system can prompt
the user to update his/her profile the first time
he/she signs in online. Thus, there is no need to
offend the users by asking for their names if all
they want is to purchase the items and move on
without wishing to establish any relationship
with the store (even in this case, the new
interface is still able to get enough information
for the users profile).
model w884 with dials to control both

temperature and humidity as seen in Figure 4 as
their base design in which the users can turn the
knob to the left to decrease and to the right to
increase the setting. The displayed dials will be
digital and a large number will be displayed for
better readability as the user starts turning the
knob. This design is truly simple and intuitive
to use. This new design is very similar to the
NEST, however, the users do not have to push
the knob, which is not intuitive for users who
are new to the technology. In addition, the dials
give the users a reference so they can adjust
their knob turning speed with minimal attention
to details.
4.2 Designing Next Generation Thermostat

Interface
The chip-equipped credit card: credit cards are

commonly used in industrialized countries.
Secondly, the threat of credit card frauds makes
users open to adapt to a new and safer
technology.
The NEST thermostat is chosen for its modern

and simplistic digital interface.
Round 1: In this round each member is asked
to operate the NEST to control the room
temperature with different preferred settings for
common daily periods such as waking up,
leaving for work, arrive after work, going to
sleep for one-week time frame. Each member is
then asked to evaluate their experiences with
the NEST thermostat. All members in both
groups express that they like NESTs slick,
modern looking digital interface design and its
ability to learn users preference temperature
patterns. There are 7 members who have no
prior experience with the device interface that
express that they need a little time to learn how
to use NEST.
Round 2: All members are briefed on
traditional and existing thermostats, including
how to operate them. All members are asked to
design the interface for the next generation of
thermostat. It is surprising that the members
have chosen the design of Honeywell 1968
ISBN: 978-1-941968-38-32016 SDIWC
4.3 Designing the Next Generation of

Secured Credit Cards
Round 1: With similar setup to the previous

case study, the members of the two groups are
asked to use their chip-equipped credit cards to
make their daily purchases for a week. The
members are asked for their experience. All
members acknowledge that the new card takes
much longer time to process (~ 1 minute while
non-chip card take less than 5 seconds) and
they have to keep looking at the card reader
digital display in order to know when the
transaction is done. Despite these issues, all
members express that they like the chipequipped card because it makes them feel more
secure when making transactions. This is an
expected result because these credit card users
are concerned about fraud and will sacrifice
usability for extra security.
Round 2: All members are briefed on how the
chip-equipped card handles its user transaction
information. The members are asked to review
the information on the chip-equipped card and a
22
non-chip card. The members are then asked to

design the next generation of credit cards. The
members come up with two designs: (1)
modifying the card readers for user experience
and (2) redesign the card for true security.
Design for better user experience: in this
design, the members want the card reader to be
redesigned to read both the magnetic stripe and
the chip on the card in one swipe. This way, the
users do not have to swipe again whenever the
reader fails to read the chip. It is also
cumbersome that the users must check the card
reader capability and to remember which card
to swipe and which card to insert for every
transaction. Moreover, the card reader should
give an easy recognizable signal such as light
up green on processing and light off on
done. These modifications help to improve
user experience.
Design for true security: this design addresses
the security of the transaction and the users
experience. In this design, the card is similar to
the non-chip card; however, there is no name,
credit card numbers, nor signature. There is
only an encrypted alpha-numeric sequence of
characters that can be decrypted by the credit
card processing center. When the user makes a
transaction, the point of sale computer terminal
will receive an image and the name of the
actual owner of the card so the salesperson can
verify the identity of the buyer. For each
transaction, the user will be asked to provide
the pre-registered information such as birthdate,
pin, billing zip code, phone number, income
range, driver license expiration date, etc.,
selected at random. This information is
currently and normally available to the credit
card issuers. Thus, with this design, the card
can be used securely for both online and in
store purchases because the users personal
information is no longer on the card.
4.4 Evaluation of the Next Generation
Interfaces
ISBN: 978-1-941968-38-32016 SDIWC
The members who participated in our study are

college students that are interested in
technology and computing, and majority of
them have no prior experience on interface
design. However, their designs are simple using
existing technology, and resemble existing
interfaces.
We evaluate the new interfaces with 10 new
members who have not participated in any
previous activities. We use the following
evaluation criteria: functionality, users
perception, learnability, interface complexity,
and task completion time.
The functionality is simple to evaluate because
the new interface must carry the same
functionality as the existing ones; The task
completion time is simply the time the users
take to complete an assigned task. Learnability
is reflected by the number of mistakes and the
amount of time the users take to complete the
task. Users perception is users preference on
which interface they would rather use after
trying them both.
For interface complexity, we use Tullis
simplified screen complexity method [15], to
evaluate the new interfaces against the
corresponding existing interfaces. In this
method, the components on each interface is
organized into rows and columns by their
alignments; the sum of all rows, columns, and
the components in each row and column are the
complexity score.
With these setups, the result on the new
interfaces are all better than the existing ones.
Especially, the check-out web-form redesigned
for Walmart.com reduces almost 80 percent of
required information and total work the users
have to perform and still manage to have the
same functionality.
In term of handling fundamental functionality
and operations of the thermostat, the new
design is more user friendly. All other criteria
23
are mostly comparable with the existing NEST

thermostat.
The credit card designs are also better than the
one that have been designed by large credit card
issuers. The credit card first design improves
the user experience by cutting down on
different tasks the user may have to do. And the
second design is truly exceptional. It provides
the best security and flexibility for credit card
users by modifying the existing interface of the
existing software that is currently used for nonchip cards. The new design is more secure and
simpler than the existing chip-equipped cards.
5 CONCLUSION
Designers should study the evolution of
interfaces designed for similar tasks to find the
pattern of a good interface design. A new user
interface design should maximize the capability
of the technology to provide the users with
greater functionality and minimal impact on
how the problem is being solved. Through our
case studies, we have shown that there is a
pattern of successful interface design across the
evolution of many different problem domains;
and that a successful design pattern often exists
in the primitive classical interface designed for
the problem.
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Computer Interaction, Fifth Edition, AddisonWesley Publ. Co., Reading, MA, 2010
P. V. Schaik and J. Ling. The role of context in
perceptions of the aesthetics of web pages over
time. International Journal of HumanComputer
Studies, 79-89, 2009
K. Holtzblatt and H. R. Beyer. Contextual design:
The Encyclopedia of Human-Computer Interaction,
2nd Ed. Interaction Design Foundation, 2014
Rory Carroll. "Bell did not invent telephone, US
rules" The Guardian, June 17, 2002
Allison Bailes. If You Think Thermostat Setbacks
Don't Save Energy, You're Wrong!", 2012
[http://www.energyvanguard.com/blog-buildingscience-HERS-BPI/bid/50152/If-You-ThinkThermostat-Setbacks-Don-t-Save-Energy-You-reWrong.]
Emily S. Gerson. Pre-plastic credit: Charge plates,
coins,
celluloids,
(November
7,
2007).
[http://www.creditcards.com/credit-cardnews/credit-collectible-coins-charge-plate-1264.php]
J. Yan, and A. S. El Ahmad. Usability of
CAPTCHAs or usability issues in CAPTCHA
design, pp. 44-52, ACM, 2008
T. Yamamoto, T. Suzuki, and M. Nishigaki. A
proposal of four-panel cartoon CAPTCHA: The
Concept, pp. 575-578, IEEE, 2010
Hossain, M., Nguyen, K. D.andRahmanM.A.
"Labeled-Image CAPTCHA: concept of a secured
and universal useful CAPTCHA", proceedings of the
4th International UBT International Conference on
Computer Science and Communication Engineering
(IC-CSCE)- pp.102-108, 2015
N. Summers. Vicarious claims its AI software can
crack up to 90% of CAPTCHAs offered by Google,
Yahoo and PayPal, vol. 2014, no. October 25, pp.
3, 2013
http://beta.fortune.com/global500, July 20, 2016
T. S. Tullis. A system for evaluating screen
formats: Research and application. Advances in
Human-Computer Interaction, 214-286, 1998
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
Donald A. Norman. The Invisible Computer,1998,

[http://leefrank.com/inquiry/booknotes/book159.htm
l]
Golden Krishna. The Best Interface is no Interface:
the simple path to brilliant technology(voices that
matter), Amazon, 2015
"Bellsouth, IBM Unveil Personal Communicator
Phone". Mobile Phone News. November 8, 1993.
ISSN 0737-5077.
J. Laugesen and Y. Yuan, "What Factors
Contributed to the Success of Apple's iPhone?",
Proceedings of the 2010 Ninth International
Conference on Mobile Business / 2010 Ninth Global
Mobility Roundtable, pp. 91-99, 2010
R. Molich, and J. Nielsen. Improving a humancomputer dialogue, Communications of the ACM
33, 3 (March), 338-348, 1990
B. Shneiderman and C. Plaisant. Designing the
User Interface: Strategies for Effective Human-
ISBN: 978-1-941968-38-32016 SDIWC
24
Education in Computational Science:

Do Successful Examples Also Create Success in Education?
Milena Janakova
Department of Informatics and Mathematics, Silesian University in Opava,
School of Business Administration in Karvina, Karvina, Czech Republic
mija@opf.slu.cz, maninformatics@gmail.com
ABSTRACT
This paper is dedicated to improving education in
computational science with support of examples
from practice based on links from social networks
as Twitter and LinkedIn. We have many of
successful examples, but the question is what the
benefits are for education. The aim is to contribute
to improve study results and optimal orientation of
students in the CRM (Customer Relationship
Management) with regard to practice. The existing
controversy
creates
inconsistency
between
a successful example in practice and education that
is not always successful. The method of solution
uses the multidimensional approach, and other
sources of information are realised surveys in CRM
courses. Achieved results show that examples from
practice are adequately supported with suitably
verified links to Twitter and LinkedIn. They
minimise the number of students with a negative
perception of courses. Benefit is good mirroring
dynamic changes in IT (information technology)
field and diversity of adopted solutions. The
challenge is to work with a growing group of
passive students and their influence through
examples from practice. Future work should focus
on increasing the interest of students in CRM and
their curiosity about appropriate methods.
KEYWORDS
Computational science, CRM, education, examples
of practice, information technology.
1 INTRODUCTION
Education is one of the oldest disciplines in the
world. People must learn new knowledge and
ISBN: 978-1-941968-38-32016 SDIWC
skills all the time. Of course, the environment

and tools are different, but motivation is the
same the need to survive and obtain benefits
that others cannot get. In a modern global
society, survival is the ability of firms,
organisations and individuals to withstand the
pressure of competition and gain a satisfied
customer. [1] In this topic, CRM brings good
examples. The aim of CRM is to care for
customers and their satisfaction and loyalty
with active support of the best market position
and profit. [2]
Achieved profits and costs, available assets and
liabilities are default terms in all companies and
organisations.
Small
and
middle-sized
entrepreneurs have similar external working
conditions as other opulent and big companies,
but their internal background is different. [3]
They work in small scales, and they must have
optimal education for all executed activities.
The work is harder with regard to the rapid
development in all fields of human activity.
CRM and IT are also rapidly developing fields.
What is positive is that IT may help to remove
market barriers and create an optimal base for
work with information. The volume of
information is growing [4], and optimal
decision-making often relies on business
intelligence and big data [5]. There is no room
for estimates and uncertainty. On the other
hand, the global information society brings a lot
of instability and unexpected situations. These
great problems create space for economic
collapse with links to unemployment and social
25
matters. [6] These problems are only a fragment

of the actual difficulties. A good defence
against most issues is sustainable development,
which helps in balancing socio-economic
activities. [7]
The present is characterized by pressure on
innovation and diversity [8], because
innovation is important for the optimal search
of new products and services. Customers must
demand such products and services. Customers
today are different from the customers of the
last century. There is the Internet and customers
share opinions about products and services.
They also search for advice and inspiration
based on experiences around the world. This
behaviour has a character of collective intelligence,
as for example swarm intelligence. [9]
In this situation, we need optimal skills,
orientation, quick information support, and the
ability to move data into information,
information into knowledge, and knowledge
into optimal decision-making for effective
business plans. [10] A large volume of data and
spreading required knowledge is rapidly
changing and both complicate the situation.
Education encounters similar difficulties.
Learning must offer complex education with
quick orientation in a topic and many links to
other areas. For many teachers, one favourite
activity is providing examples from practice,
led by good examples from practice.
2 CONTROVERSIES IN EXAMPLES
OF PRACTICE
Examples from practice, good examples from
practice, examples of best practice, good
practice examples, examples with solutions,
best practices, or examples for practice this is
only a fragment of the terms that are used for
accessing experiences from realised activities.
Working teams want to publish achieved results
as well as inform actual or potential customers
about new opportunities. Such a situation is
visible in CRM. There are many positive
examples on the Internet with a summary of
ISBN: 978-1-941968-38-32016 SDIWC
available functions and activities. Some

examples are:
Best practices for developing with
Microsoft Dynamics CRM. [11]
The five biggest benefits of CRM
systems. [12]
12 Amazing CRM charts you dont
want to miss. [13]
Customer service and CRM: 10 best
practices. [14]
From older studies and practices, interesting
sources are:
Gartners top 54 CRM case studies. [15]
Best practices: customer relationship
management. [16]
At first glance, all CRM projects are good and
successful. However, this is not true; there are
controversies. If we look at projects from the
field of IT, we see problematic results and
experiences from information technology.
Some users prefer using MS Office (MS Excel)
over specialised software because this software
is too complicated. [17] MS Office may
certainly bring basic information in a given area
of interest, but suitable results are created by
software that is more sophisticated. Another
problem is the time delay of realised work on
an IT project. There are projects that are not
finished in time. [18] We must also remember
that IT users automatically need to work with
data. They prefer to work with software based
on their own skills according to their personal
preferences. One of the great difficulties is low
orientation in a user interface for effective
switching to other software and adaptation of
an optimal method of work in a given situation.
3 RECOMMENDED METHOD
FOR SOLUTION
The active solution to the existing controversies
is based on objects [19] and the
multidimensional approach [20] with support
surveys. Objects are one of the default entities
for the description of actual reality. The benefit
of this approach relies on the synergy between
26
data and methods of manipulation with them.

The multidimensional view has its foundation
in methodologies for the design of information
systems such as MDIS, Multidimensional
Development of Information System. [21]
Now, this view is important with regard to data
science and various forms of intelligence
(artificial intelligence, business intelligence,
customer intelligence, data intelligence). The
reason is that data and information have to be
examined from different perspectives. Objects
of interest are examples from practice with
links to study activities and a students
perception of a course.
Successful examples from practice also bring
success into education. What influence do
negative examples from practice have? Are
these examples important or do they bring
negative motivation into education? Students
have to know a new skill in a selected topic,
and new knowledge is connected with errors,
mistakes, and a lack of understanding of the
relevant procedure. Common examples from
lectures and practical examples from seminars
improve understanding.
Examples from practice have a similar
influence. Examples from practice (positive or
negative) may improve orientation in a user
interface and selection of adopted methods.
From this point of view, negative examples
may prevent similar mistakes from happening
and protect IT users against disenchantment
with the obtained results. The students views
on realised activities in the CRM course bring
more information on this topic.
4 PRACTICAL EXPERIENCES
FROM THE CRM COURSES
Database
Technology
for
Support
Management and CRM (a CRM course). This
CRM course is dedicated to students in the
continuing segment of the Masters study
programme. The aim of the course is to teach
students the characteristics of CRM systems
and their philosophy.
Customer relationship management uses
a modular system based on needed applications.
Base data is oriented on records about business
partners and customers. In this course, students
learn the detailed explanation of a users
environment, the analysis of customer contact
with the company, the question of security and
implementation. There are also links to
knowledge
management
and
business
intelligence. Practical examples play an
important role too.
Surveys say more about a students view of this
course. These surveys ask about satisfaction
with this course, benefit from course structure,
missed topic, and other suitable links or advice
according to student preferences. The important
questions are:
Does
the
course
fulfil
your
expectations?
Is the course structure suitable?
Evaluate innovation in the course (links
to Twitter or LinkedIn).
Evaluate topic areas and their
suitability.
Do you have a recommendation about
innovations for this course?
By comparing answers in academic year
2013/2014 (with 15 students) and academic
year 2015/2016 (with 17 students), we gain
interesting information. Please see Table 1.
This academic year 2015/2016, the School of

Business Administration in Karvina (Silesian
University in Opava) opened a course called
ISBN: 978-1-941968-38-32016 SDIWC
27
Table 1: Student answers to questions in percentages in academic year 2013/2014 and 2015/2016.
Answers
Questions
Does the CRM
course fulfil your
expectations?
Is the course
structure suitable?
Do you have
a recommendation
about innovations
for this course?
Yes
Rather Yes
I do not know
Rather No
No
academic year 2013/2014 (2015/2016)

56% (25%)
22% (50%)
11% (25%)
11% (0%)
0% (0%)
33% (50%)
56% (25%)
11% (25%)
0% (0%)
0% (0%)
0% (13%)
22% (0%)
34% (37%)
33% (37%)
11% (13%)
In 2013/2014, about 78% of students (with

answers I do not know, Rather No, No) did not
supply any advice based on the question Do
you have a recommendation about innovations
for this course? About 22% of students (with
answer Rather yes) advised adding links to
available jobs. In 2015/2016, students evaluated
links to Twitter and LinkedIn. It is surprising
that about 37% of students (with answer I do
not know) have no interest about innovations in
CRM course, but interest about innovations is
always apparent in all surveys. Students
advised adding more information about singleuser applications and MS Dynamics CRM
(2013/2014), or MS Outlook and database
structure with links to CRM (2015/2016).
If all positive answers (Yes, Rather Yes) and all
negatives (Rather No, No) are grouped for
question about expectations with the CRM
course, there are only three values for course
evaluation through the variables Positive, Do
not know, Negative.
The following figure shows changes in the
evaluation of the CRM course.
Please see Figure 1.
ISBN: 978-1-941968-38-32016 SDIWC
80
70
60
50
40
30
20
10
0
78 75
25
11
Positive
Do not
know
2013/2014
11 2015/2016
0
Negative
Figure 1: Diversification of the overall perception of the

course from positive to negative in percentages.
From the teachers point of view, the positive

fact is that there are zero students with
a negative perception. The important matter is
also selection of suitable links from Twitter and
LinkedIn. The added value brings accounts of
well-known IT companies, verified users with
long-term interest about given topic as
distributors, partners and key users of IT
products or journals. For example, between
such accounts belong @CRM, @microsoftcrm,
@oracleCRM, and @SugarCRM.
28
Some students have difficulties with a volume

of information and its diversity. They need
more time for orientation and practical work.
A big problem is the greater group of students
that do not know. They do not think about
such topics and they passively receive
education.
5 RESULTS AND DISCUSSION
In 2013/2014, the CRM course contained
standard lectures (PDF files) and a lot of
exercises for tasks in seminars. Presented
examples described positive results. Available
examples were oriented on the practical use of
selected products such as:
eWay CRM [22],

Facebook [23],
IZIO [24],
SugarCRM [25],
QlikSense Desktop [26] to demonstrate
the link to business intelligence.
Based on student recommendations, the CRM

course offered thematic areas for examinations.
There is also created place for actual
information from Twitter and LinkedIn. In
2015/2016, these innovations were added, but
positives and negatives still exist.
Please see Table 2.
Table 2: Achieved positives and negatives in the CRM course in 2015/2016 compared to 2013/2014.
Positives
Negatives
Expectation from the course is filled

(there is no negative evaluation).
More level of passivity

(more students do not know).
Course structure is suitable

(there is no negative evaluation).
Low level of interest about topic structure

Still existing interest about innovations

(some students say yes).
Lower interest about innovations

There are no negative expectations from course

and course structure. The challenge is working
with 25% of students that do not know. Maybe
they have no interest about this topic, or they
are more passive in education. They only wish
to successfully pass the examination. A large
group of students does not have interest about
innovations for the CRM course (about 37%).
One solution is to also bring negative examples
from practice and invite students to minimise
existing mistakes through their own solution.
This method starts a discussion on the given
topic.
Student satisfaction from lectures in realised
courses is the centre of focus for a lot of
analyses, surveys, evaluations or feedback with
ISBN: 978-1-941968-38-32016 SDIWC
links to teacher education and scientific

outcomes. From interesting analyses, there are:
Educating the educators involved in
homebound training and work inclusion.
[27]
Feedback improvement of question
objects. [28]
Universities have limited impact on
students soft skills development.
[29]
Many of them offer analyses from the views of
students and teachers. The aim is to know the
preferences and needs of customers in
education. There are students that will fill
existing demands from employers, firms and
29
organisations. Many of them will have their

own enterprise. In any case, all must face
unexpected conditions and the hard market
environment, and all wish to succeed.
Visualisation and simulations increase interest
in education and help students understand
a given topic. The question is about negative
examples from practice. Based on the above
data from a CRM course, it is clear that positive
examples from practice are good, but there are
reservations. Special interest must be paid to
examples with zero power in practice. The
initial intent focused on increased interest in the
CRM course and needed skills for optimal work
may fail. The reason is doubt from students that
they do not know the suitable methods, and that
they will not solve the assigned tasks. This is
the challenge for future work in the CRM
course, which will aim to increase positive
expectations and skills for resolving default
tasks in CRM.
Twitter and LinkedIn. Such study activities

help to better perceive the CRM course and its
structure, but there are still passive students.
They do not think about innovations or search
for optimal work methods in selected software.
From this point of view, successful examples
are not a unique way to succeed in education.
There are examples with negative and zero
power in practice. It will be interesting to track
student reactions to them in future work.
REFERENCES
[1]
[2]
[3]
[4]
6 CONCLUSION
Education in computational science is
sophisticated activity that requires interaction
between all participants such as students,
teachers, firms and organisations from practice.
Information is available via the Internet; there
are many examples from practice mentioned to
achieve a better understanding of selected
topics. The global information society offers
many possibilities and unexpected connections.
In this reality, unfortunately, controversies still
exist from areas of innovation, diversity with
links to economic collapse and socio-economic
activities. Complex education is needed with all
examples from practice that describe an active
approach for the solution of required tasks.
CRM systems are also in the centre of interest.
Usually, CRM lectures offer needed study
documents and practical examples in available
software products. These study documents and
activities help to get better skills for practical
work based on practical examples and thematic
areas. Through the realised surveys, students
positively evaluate these examples with links to
ISBN: 978-1-941968-38-32016 SDIWC
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
P. Flory, The Complete Customer Relationship

Management (CRM). Handbook, UK: Directory of
Social Change. ISBN-10: 1906294720, 2013.
G. Slater, How To Implement A CRM System for
Maximum Profit: A Guide For Business Owners
Who Want To Profit From Customer Relationship
Management. Slater Trading Pty Ltd. ASIN:
B00CAEDFOM, 2012.
K. Krivkova, Small and medium-sized enterprises the backbone of the economy. [online] Available at
<http://sophispol.webnode.cz/male-a-strednipodniky-pateri-ekonomiky/>, 2011.
R. Alleyne, Welcome to the information age 174
newspapers a day. [online] Available at
<http://www.telegraph.co.uk/science/sciencenews/8316534/Welcome-to-the-information-age174-newspapers-a-day.html>, 2011.
C. Howson, Successful Business Intelligence:
Unlock the Value of BI & Big Data. USA:
McGraw-Hill
Osborne
Media.
ISBN-10:
007180918X, 2013.
A. Walker, Eurozone unemployment reaches new
high, News Business. [online] Available
at<http://www.bbc.co.uk/news/business-20943292>,
2013.
T. Baumgartner, T.R. Burns, and P. DeVille, The
Shaping of Socio-Economic Systems: The
application of the theory of actor-system dynamics
to conflict, social power, and institutional
innovation (Routledge Library Editions: Social
Theory). Routledge. ISBN-10: 1138989991, 2016.
D. Livermore, Driven by Difference: How Great
Companies Fuel Innovation Through Diversity.
AMACOM. ISBN-10: 0814436536, 2016.
H. Dresner, Wisdom of Crowds Business
Intelligence
Market
Study.
CreateSpace
Independent Publishing Platform. ISBN-10:
1478270322, 2012.
R. Knudson, Building Business with CRM. CO,
USA: We Speak You Learn, LLC. ISBN-10:
0981511848, 2012.
Microsoft Dynamics CRM, [online] Available at
<https://msdn.microsoft.com/en-us/library/
gg509027.aspx>, 2016.
P. Robles, The five biggest benefits of CRM
systems.
[online]
Available
at
<https://econsultancy.com/blog/66287-the-fivebiggest-benefits-of-crm-systems/>, 2015.
30
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
S. MacDonald, Amazing CRM Charts You Dont

Want
To
Miss.
[online]
Available
at
<http://www.superoffice.com/blog/crm-charts/>,
2014.
G. Diorio, Customer Service and CRM: 10 Best
Practices.
[online]
Available
at
<https://www.cloudsherpas.com/partnersalesforce/customer-service-and-crm-10-bestpractices/>, 2013.
E. Thompson, Gartners Top 54 CRM Case Studies,
Sorted by Industry, for 2005. [online] Available at
<https://www.gartner.com/doc/481170/gartners-top-crm-case>, 2005.
I. Gordon, Best Practices: Customer Relationship
Management.
[online]
Available
at
<http://iveybusinessjournal.com/publication/bestpractices-customer-relationship-management/>,
2002.
InvestIntech, Microsoft Excel Experts Predict The
Future Of Excel In Business Intelligence. [online]
Available
at
<http://www.investintech.com/
resources/blog/archives/5718-experts-predict-thefuture-of-excel-in-business-intelligence.html>, 2015.
L. McQuerrey, What Causes a Project Delay?
[online] Available at <http://yourbusiness.az
central.com/causes-project-delay-16867.html>.
B. Dathan, and S. Ramnath, Object-Oriented
Analysis, Design and Implementation: An Integrated
Approach (Undergraduate Topics in Computer
Science). Springer. ISBN-10: 3319242784, 2015.
S. Alkire, J. Foster, S. Seth, M.E. Santos, J.M.
Roche, and P. Ballon, Multidimensional Poverty
Measurement and Analysis. Oxford University
Press. ISBN-10: 0199689490, 2015.
M. Kekovsk, M. Drdla, Strategy management of
business information. Praha: C. H. Beck. ISBN: 807179-730-8, 2003.
eWay
CRM,
[online]
Available
at
<https://www.eway-crm.com/>, 2016.
Facebook,
[online]
Available
at
<https://www.facebook.com/>, 2016.
IZIO, [online] Available at <https://www.izio.cz/>,
2016.
SugarCRM,
[online]
Available
at
<https://www.sugarcrm.com/>, 2016.
Qlik Sense Desktop, [online] Available at
<www.qlik.com/products/qlik-sense/desktop>, 2016.
G. Trentin, F. Ravicchio, M. Repetto, Educating the
educators involved in homebound training and work
inclusion. International Conference on Education
Technologies and Computers, ICETC 2014, Lodz,
Poland, September 2014, pp. 36-43. ISBN: 978147996248-8.
M. Hruby, Feedback improvement of question
objects. International Journal of Continuing
Engineering Education and Life-Long Learning, vol.
26, Issue 2, April 2016, pp. 183-195. ISSN:
15604624.
UK Engagement Survey: universities have limited
impact on students soft skill development, [online]
Available
at<https://www.timeshighereducation.com/news/ukengagement-survey-universities-have-limitedimpact-students-soft-skill-development>, 2015.
ISBN: 978-1-941968-38-32016 SDIWC
31
Projects for Teaching Algorithmization in Primary Schools

Hedviga Petrukov, Milan Lehotsk, Janka Majherov and Igor ernk
Department of Informatics
Faculty of Education Catholic University in Ruomberok
Hrabovsk cesta 1,034 01 Ruomberok Slovakia
E-mail: hedviga.palasthy@ku.sk, milan.lehotsky@ku.sk, janka.majherova@ku.sk, igor.cernak@ku.sk
ABSTRACT
In this article the current trends in teaching
algorithm development and programming in
elementary schools in Slovakia are presented. The
first part is devoted to the subject Informatics and
describes learning environments suitable for
primary school students. The second part deals
with the methodological aspects of teaching
algorithmization and the appropriate programming
environment. We describe an educative
application containing the appropriate tasks for
teaching of algorithmization suitable for pupils in
a primary school.
KEYWORDS
teaching algorithm, programming at primary
schools, children's programming languages
INTRODUCTION
A content of the subject informatics has

developed in relation to the contemporary
state of technology. A great shift has occurred
in types of tasks and problems that have been
dealt with in different periods. Initially the
subject informatics was identical to
programming. Later, as programming in
schools was suppressed, informatics has
become teaching of applications. Now once
again programming is a part of informatics
education.
The
Algorithmization
and
Programming status is also supported by the
State educational program, which contains
thematic unit procedures, problem solving
and algorithmic thinking.
Similarly to mathematics, the subject
informatics teaches students how to think
logically and follow steps that are described
by an algorithm. People encounter the
ISBN: 978-1-941968-38-32016 SDIWC
algorithms in their daily activities although

only a few realize it. While doing some of
activities, people follow a given action,
perform the undertaking and therefore in fact
act like programmers. It is appropriate to
familiarize students with such situations. By
means of understanding of a correct sequence
of steps leading to a desired action, they
develop logical thinking and thus learn how to
deal with complex situations in life in a clever
way and without a need of great effort.
2
ALGORITHMIZATION IN
INFORMATICS IN A PRIMARY
SCHOOL
The educational content of informatics in the

State educational program for primary schools
is divided into five thematic areas. The topic
Algorithmic problem solving introduces
students
to
specific
troubleshooting
procedures through ICT. The students become
familiar with concepts such as algorithms,
software, and programming. The biggest
benefit of this domain is that the students
acquire basic algorithmic thinking and an
ability to think about problem solving using
ICT. They learn to reflect on various
parameters of efficiency in different solutions
to problems. At the same time, the students
learn different procedures and mechanisms
that can be used to solve problems from
diverse fields of science [1].
The subject informatics is compulsory in
primary education since 2008. According to
the innovative curriculum of lower secondary
education from 2015, the number of teaching
hours is one hour per week at the primary
level in the third and fourth year [2]. At the
secondary level, informatics is taught from
32
the fifth to eight grade also at the rate one

hour per week. As a school year consists of
33 weeks, it is a total of approximately 120
hours of informatics. Programming can be
planned for about a quarter of teaching hours
of Informatics, which yields about 30 hours at
the secondary level.
The teaching hours count of informatics (and
programming) can increase up to twofold, if a
school decides to designate free hours to
strengthen this subject. An extent of
Informatics is therefore the first key
parameter in the planning of teaching
programming. The second parameter is an
allocation of content to classes. During a
planning of the informatics content the
"spiral" approach is used. This means that
thematic units of the school course are
repeated in multiple editions always at an
increasing level. A time lag between
teachings of various parts of the theme in
different grades makes room for a
consolidation of knowledge.
Although the state educational program
introduces teaching of algorithmization
already at the primary level, in many primary
schools a move to the second level of lower
secondary education or to the programming is
omitted. The teachers often do not have
enough experience with teaching of
programming or they lack methodological
materials.
In years 2006-2015, we conducted a survey
where we measured the number of hours of
algorithmization and programming teaching
at primary schools. In the reporting period
teachers devoted in the first and second grade
of primary schools respectively 17% and 20%
of the total number of teaching hours in the
subject informatics for topic area Procedures,
problem solving, algorithmic thinking.
Regarding programming languages, teachers
prefer the language Baltie in the primary level
(53% of teachers), and the platform Imagine
in the secondary level (83% of teachers).
Surprisingly, only few teachers were active in
preparing students for competitions in
programming (only 5.6% of teachers in the
secondary level of primary school). The
longest time was dedicated by teachers to
ISBN: 978-1-941968-38-32016 SDIWC
work with text, graphics and presentations.

On the other hand, programming was featured
at 14% of lessons only [3].
The pupils in elementary school are able to
work in children's programming languages, so
we do not have to deny them this experience.
The pupils should be made familiar with
programming using games and suitable
programming languages. Nowadays there is a
good selection of various children's
programming languages.
3
LEARNING ENVIRONMENTS FOR

ALGORITHMIZATION
Motivating pupils in a primary school to study

programming and teaching them logical and
algorithmic thinking requires "special"
programs that are attractive and manageable
for children of corresponding ages. These
"special" programs are called children's
programming
languages
or
small
programming languages or microworlds.
Their visual environment is intuitive and easy
to operate.[7]
The childrens (or small) programming
languages are a combination of an object and
a language. The language consists of simple
commands by means of which a student
controls the object (e.g. a turtle robot,
magician or other entity) in the microworld.
The advantage of these languages is that the
student views on-screen behavior of the
controlled object. They learn basic
programming concepts and understand the
more complex notions of the structured
programming. The commands are easy to
remember for the pupils.
The teachers of informatics can choose from
multiple children's programming languages,
such as Imagine, Baltie, scratch or Kodu. We
describe them individually in the subsequent
part of the article.
3.1 Imagine
Imagine was created in 2001 as an indirect
successor to the Comenius Logo program [4].
It is an object language driven by events. "The
hero" (a main character) is a turtle which can
33
change its shape and draw within the area.

The aim of authors was to create a modern
programming environment that is easy to
control for an experienced user and also for an
absolute beginner a pupil in the primary and
secondary level of an elementary school. The
advantage for younger pupils is that they do
not have to write a coherent program that
focuses on the solutions of all possible
situations. Instead they could write many
small sub-programs, sometimes referred to as
one statement, which will be called
automatically during various events, for
example a mouse click, drawing, collision
with other objects, etc. It is possible to define
different subroutines for various objects and
their events. The environment has an
integrated
simple
editor,
background
graphics, multimedia, and an option to
connect to the Internet (Figure 1).
An interesting new feature is a possibility to
publish projects on the Web: using the
Imagine plug-in, we can run the finished
projects from the network in an Internet
browser. A bitmap editor is also a component
of the Imagine. It is used to prepare images
especially for the LogoMotion shapes of
turtles, but also to create wonderful
animations. Imagine has two language
variations - Slovak and English.
Figure 1: Imagine
3.2 Scratch
Scratch is an environment that was created in
2007 originally for children six to sixteen
years old. In this program they have an
opportunity to discover, experiment and
create. It was developed in the MIT Media
ISBN: 978-1-941968-38-32016 SDIWC
Lab Lifelong Kindergarten by a team led by

Mitchel Resnick[6]. This team has developed
multiple new technologies with which people
can create, invent new things and learn in a
playful and technically simple manner.
Initially, Scratch was created under the idea of
environmental image, but this has been
replaced by a concept of commands being
called by dragging and cards being dragged to
the commands [5].
Scratch is an environment suitable for
teaching programming. It can be used easily
to create interactive stories, animations, or
games, and pupils can share these on the
global website.
The environment in Scratch consists of four
areas, a main menu and several auxiliary
action buttons. The main menu is userintuitive and consists of the same parts as
many other menus. The main priority when
creating this new environment and language
was that its use should be as intuitive as
possible and easy to be learned by children
without previous programming experience.
The environment (Figure 2) is divided into
several parts. On the left-hand side there are
cards with various commands. The commands
are grouped into eight categories labeled by
colors: Motion, Sight, Sound, Pen, Control,
Perception, Operators and Variables. In the
center there is information about the currently
controlled character, its properties and Scripts
(scenarios). In the bottom, there is a list of all
the characters. Finally, at the very top there is
the main menu where you can, for example,
change the language, or open, save, and
publish projects. Controlling of the
environment is done by moving cards with the
commands from the Palette to the Script, and
those relate to a particular character which is
currently in the Scene. In the Script part, the
cards can be combined into blocks and
arbitrarily reshuffled.
The cards are designed to fit together in the
correct way. It is not possible to make a
syntax error, as the environment does not
allow it.
34
experiences of with Scratch, all in different

languages.
Figure 2 Environment Scratch 2.0
The three basic principles of the Scratch

environment are Imagine, Program, Share.
[4] A website connecting a Scratch online
community (see Figure 3) has been created to
share projects created in Scratch, as well as
other materials and ideas related to this
program.
Figure 3 Home Scratch with user login (language Czech)
This page motivates children to program and

create in the environment of Scratch. Here
they can find many ideas and inspirations.
The users communicate with each other
through comments on a project, or via forums.
The source code for each project can be
downloaded directly from the site. Projects
can be published directly on the website and
run by the Java applet player known as
Scratch player. Moreover, projects can be
easily downloaded or opened directly in
Scratch.
In July 2009, a website (Figure 4) was
launched where teachers and parents can
share their experiences with Scratch, share
materials, and ask for advice. [5] This website
contains materials of various kind, stories,
documents,
discussion,
and
teachers
ISBN: 978-1-941968-38-32016 SDIWC
Figure 4 Page Scratched associating teachers who use

Scratch
3.3 Kodu
Another suitable environment for the teaching
of algorithms in elementary school is the
environment Kodu [7]. Kodu is a free
program that allows you to create games on
the computer. It uses a simple visual
programming language. It includes own
library of objects and allows the easy visual
programming without knowing any code. The
pupils present their creativity to solve
problems and at the same time they program
in a simple form.
Kodu
contains
interesting
graphical
environment and the programming in it does
not require the knowledge of programming
languages and the creator monitors his
progress in real time. The development
environment is designed for educational
purposes and uses a library of 3D models. The
code is composed of the commands for the
object. The toolbar includes the creation of
buildings, roads and surfaces. It contains a
large library of the materials. The pupils can
change the sky, lighting, camera settings,
transparency, wind and other effects in the
world. They determine how the game begins
for instance: the name of the world,
description of the world, countdown. In
addition to setting the world it is possible to
set each object separately.
The program Kodu in the operating system
Windows 8 and higher is installed using the
Windows Store. Just enter a search Kodu
Game Lab. The program supports multiple
languages. When using Windows 7, we need
35
to download the installer from the official

website Kodu.
Figure 7 Rules, the development of logical thinking
Figure 5 Program Kodu
THE APPLICATION FOR

TEACHING ALGORITHMIZATION
The application Imagine Turtle Imagine is

designed for elementary school students by
studen. It is divided into five themes which
interlock. It is possible to work with them in
one grade or it can be divided into more
grades. The tasks of the individual topics are
designed to match the knowledge of the
pupils of given age. The separation into
individual topics is made according to the
requirements of the state education program
for science curriculum [9].
The role of pupils in the task Ordering articles

is that objects are sorted correctly according
to certain rules. The main scope of these tasks
is that the pupils learn to sort objects
according to given rules. The pupils move
objects in the tasks according to the rules in
the space provided in the application. The
object is moved back to the original position
in the case of the erroneous placement of it.
The aim of the task Composing by pattern
is to create an image by pattern. The task
includes seven variants of pictures. This task
does not include feedback to pupils and it is
a goal check the accuracy of solutions. The
difficulty of the images increases with each
variation.
The responsibility of students in the task
What continues is correctly fill the given
sequence. There are four options each of
which comprises a sequence of images.
The second theme Instructions, process,
solution according to the instructions should
be used when pupils are familiarizing with
instructions, create their own tutorials
a series of steps, follow the instructions at
work and continue to develop their logical
thinking.
Figure 6 Application Turtle Imagine
It is possible to use the first topic Rules - the

development of logical thinking. The pupils
are familiarized with the rules and understand
their meaning. The pupils develop their
logical through activities and tasks of the
tutorial Turtle Imagine. There are three kinds
of tasks: Sorting subjects, Composition by
a master and What is going on (Figure 7).
ISBN: 978-1-941968-38-32016 SDIWC
Figure 8 Sample of the task Sort
36
The task Sort includes three variants. In the

first variant (Figure 8) the task is to create
a sequence of steps for the proper formation
of origami. The second variant requires
creating a manual for cooking spaghetti. The
task of the third option is to create the correct
sequence of steps of binding ties.
The main scope of these challenges is that the
students learn to create instructions
encountered in everyday life. The pupils in
these tasks move objects to a specified
location and thereby create instructions for
given activity. In the case of an incorrect
placement the object is returned to the
original position.
By third theme Introduction to Programming
the pupils are confronted first time with the
programming. They become familiar with a
simple and fun programming while using the
tutorial tasks Turtle Imagine. The main
objective is to promote and develop pupils'
logical thinking. A part of this theme
familiarizes with the environment Imagine
and with basic turtle commands.
The theme contains two types of tasks:
Labyrinth and Wash the floor. The goal is to
navigate properly the turtle into the house.
The role Wash the floor (Figure 9) is to
collect garbage off the floor. The pupils
navigate the turtle using the buttons.
There are the tasks Bring a turtle into the

house and drawing pictures in the theme
Drawing in Imagine.
The role of the pupils is to orientate the turtle
to the house through commands (forward,
backward, left, right). The button Another
option restarts the task with various
placement of house and turtle. The button
From the start restarts the job with the initial
displacement of houses and turtles.
In the tasks Drawing images II (Figure 10) is
to draw submitted images with the use of
learned commands. The pupils draw the
pictures to the right of the screen. They work
with the basic commands (forward, backward,
left, right) and commands to work with a pen.
Figure 10 Sample tasks Draw images II
The last topic Imagine - creating procedures is

appropriate to include into teaching after
sufficient practice the previous two issues.
The pupils become familiar with creating
custom commands through procedures.
In the task Teach turtle something new
(Figure 11) the pupils draw repeated pictures
by creating procedures.
Figure 9 Sample tasks Wash the floor
The theme Drawing in Imagine follows the

previous Introduction to programming. The
pupils practice all basic turtle commands and
they work with a pen. The task is to repaint of
the images with the use of the commands in
the programming language Imagine.
ISBN: 978-1-941968-38-32016 SDIWC
Figure 11 The task Teach turtle something new
Since the application is designed for students

of primary level, the role of the individual
topics is processed in the form of games.
37
PROJECTS FOR PROGRAMMING

IN SCRATCH
We show several projects suitable for

teaching the programming language Scratch.
The students are able to create these projects.
The user manages a spaceship with the mouse
and the ship will move with some speed in the
game Space. The asteroids with various
shapes and sizes move on the surface. Their
movement is random. The yellow stars move
on the surface similarly too. It turns randomly
and the game ends if the ship hits an asteroid.
There are flames flying from the engine of the
ship during the movement (animation). The
game can register the highest recorded score.
The task of the player is to drive the shark by

the mouse. He must catch yellow-green small
fishes and pay attention on the "toxic" spotted
fish (a score is reset).
Figure 14 Game Aquarium
Conclusion
Figure 12 Game Space
The game Pexeso consists of twenty pairs of

cards. Each card is initially moved to a
random position. The theme of Pexeso is a
castle. The game is for two players. The
player can rotate with only two cards at the
same time.
Figure 13 Game Pexeso
ISBN: 978-1-941968-38-32016 SDIWC
In our paper, we describe different examples

of using of educational software in teaching
of algorithmization in the lower secondary
level of education.
When working with primary school pupils it
is necessary to proceed in a manner other than
for the secondary school pupils. The learning
of algorithms can be seen as a more complex
topic also for older pupils. It is appropriate
algorithms to teach algorithms through play,
motivation tasks appropriate for the age and
knowledge of students for the pupils at the
primary level. The primary level pupils like to
discover new possibilities and learn to work
with new technologies.
It is important to remember that children are
curious by birth. So it's good to incite that
pupils' curiosity and also to formulate the
selected jobs.
When working in the program Imagine it is
necessary to pay attention to constantly repeat
commands. Each pupil needs its own pace for
training the commands and for creation jobs.
It is necessary to save the Turtle Imagine
Turtle tutorial on the hard drive of all
computers so the students can work at their
own pace.
If we want the student learns constructively
lets project the lessons so that pupils are
activated by the appropriate motivation. We
38
create sufficient room for experimentation

and collecting own experiences. The
appropriate roles in the zone of the following
evolution move the pupils' knowledge further.
We include to the lessons playful and creative
activities to maintain positive communication
climate.
The role of the teacher is strongly changed.
He is no more only instructor in the computer
aided education but more manager and
moderator of pupils. He must prepare
activities in which the educational software
and its programming aspects can support
construction of new knowledge by every
pupil. Solving logical problems helps develop
a logical and algorithmic thinking by pupils.
The various tasks have various levels of
difficulty and so we have to devote much time
for choosing the right tasks. They have many
advantages: a visualization of algorithms,
interactivity and a higher motivation of the
student. Our pedagogical experiences have
shown that this kind of teaching is appropriate
to students and support and develop their
creativity.
Activities with educational software can
effectively assist teachers in supporting the
pupils' cognitive process. Pupils can develop
its formal and logical reasoning, cooperation
and communication. They gain skills that are
necessary for the research work, e.g. an ability
to implement a simple research project, to
formulate a problem, to look for the solution
and cause context and to learn how to use
various methods of problem solving. The
characteristic
features
for
the
new
competencies are integration, connection and
non-demanding extension of known subject,
modeling and using of different methods by
the problem solving.
is supported by Technology Agency Czech

Republic.
REFERENCES
[1] J. Gunaga, M. Karasov, Using ICT in
Informatics Education and Selected Subjects in
Primary
Education
for
Developing
of
PupilsAbilities, In ISSEP 2014: Local Proceedings
of the 7th International Conference on Informatics in
Schools: Situation, Evolution and Perspectives,
Istanbul: Ankara University, 2014. p. 33-41.
[2] The
state
educational
program
Online:
http://www.statpedu.sk/clanky/inovovany-statnyvzdelavaci-program
[3] J. Majherov, Innovation of Curricuka in
Informatics, In DidInfo 2015, Univerzita Mateja
Bela v Banskej Bystrici, 2015. s. 191.
[4] Programming
language
http://imagine.input.sk/
Imagine
Online:
[5] Resnick Programming language Scratch Online:

https://www.media.mit.edu/people/mres
[6] Programming language Scratch
http://scratched.gse.harvard.edu/
edu
[7] Programming
language
KODU
http://www.kodugamelab.com
Online:
Online:
[8] The library of open educational materials. Online:

http://kniznica.
sospreskoly.org/home/course/content.php?_cid=345
[9] Musa S., Ziatdinov, R., Griffiths, C. (2013).
Introduction to computer animation and its possible
educational applications. In New Challenges in
Education. (pp. 177-204). Ruomberok, Verbum.
Acknowledgements
The article was prepared within the project KEGA
002UJS-4/2014 Interactive electronic learning
materials to support implementation of modern
technology in teaching mathematics and
informatics and by the project TA04031376
Research/development training methodology
aerospace specialists L410UVPE20. This project
ISBN: 978-1-941968-38-32016 SDIWC
39
Cluster Sampling for the Demand Side Management of Big Data

Yongxin Zhang Hong Shen
School of electrical & electronic engineering, North China Electric Power University
2 Beinong Road, Huilongguan Town, Changping District, Beijing, China
459368011@qq.com
ABSTRACT
In view of the DSM under the big data environment,
an improved FCM clustering is proposed, and the
daily load curve of the whole study area was obtained with the electricity data. According to the
formulation of the TOU price, which is consistent
with the characteristics of local users is given. The
electricity suggestions based on the specific user
load curve is provided, including the return of the
DR. Subsequently, the sampling division is put forward to expand the improved model. Finally, the
method is tested by the actual data, and the results
show that it has a processing speed 10 times of the
direct processing when the data is more than 10000.
KEYWORDS
big data; smart grid; data mining; power demand
side management; clustering analysis.
INTRODUCTION
Power demand side management has been

widely used in many countries in the world
in recent decades, and has achieved good
results[1]. With the development of information technology, Chinas power industry has
entered the era of big data[2]. Based on
the new electricity reform, the new method
of power demand side management under the
background of large data is studied. In smart
grid, it guides users to participate in power
grid interaction, which can not only help enterprises to save energy and reduce the cost of
electricity, but also have great significance to
the safe and economical operation of power
network and the sustainable development of
the country[3]. At present, the application of
large power data in China is mainly dependent
on the integration of power enterprise group,
three operational monitoring (control) center,
ISBN: 978-1-941968-38-32016 SDIWC
and the three major power data center. In addition, they are weak in unstructured data processing and data deep mining analysis needs to
be strengthened[4][5].
We take electricity suggestion which can be applied to the demand side management as the research object, and a solution method for large
data environment is proposed: Firstly, the matrix vector is used to improve the FCM clustering to make it conform to the processing of
the data, and to get the whole power load curve
of the study area. Then according to the rules
of TOU electricity price and the whole load
curve, we use the method of mid-point selection, and give TOU electricity price which
conforms to the local electricity habits. About
the specific users daily load curve in this region, based on the above TOU electricity price,
we through the analysis give the suggestions
on the way of electricity, and the calculation
of the demand side response returns. Secondly, we propose the idea of sampling division to improve the operational efficiency of
large-scale data processing. Finally, based on
the actual data obtained by the investigation,
the performance of the two methods (the sampling method and the direct clustering method)
are compared by the example test, and the results show that the feasibility of sampling partition clustering is proved in large data environment.
2
FCM CURVE CLUSTERING
In order to obtain the users electricity characteristics, so as to further tap the electricity value information, we consider using FCM
clustering. FCM clustering is a classical fuzzy
clustering algorithm[6], which can be used to
estimate the type of data points. Actual user
data is often power information on each time
40
node in a period of time, and we need the overall consideration of each time node in this time
period, that is to use the users power curve
as the sample unit, clustering a certain number
of user data sample unit. Therefore, this paper proposes a clustering model of FCM curve,
which is suitable for the estimation of power
consumption, and its description is as follows:
c X
n
X
J(X; M, Z) =
(ij )m ||xi zi ||2 (1)
i=1 j=1
Among them, n is the sample number of clusters, and c is the number of categories.
is a kind of fuzzy membership degree of a
certain type of samples. xi represents the
power data of the No.j sample, containing p
data. zi is a kind of cluster center, Z =
[z1 , z2 , . . . , zc ], zi Rn , m is fuzzy index.
Define the distance between the users electricity data and the cluster center:
2
Dij
= ||xj zi ||2
(2)
The updating equations of membership degree

and cluster center Z are respectively:
1
,
2/(m1)
r=1 (Dij /Drj )
1 i c, 1 j n (3)
Pn
m
j=1 ij xj
z i = Pn
,1 i c
(4)
m
j=1 ij
ij = Pc
After Setting the target function precision ,

fuzzy index m, maximum number of iterations Tm , and membership degree , by (4)
we initialize the fuzzy clustering center Z, and
reuse (3) and (4) iteratively to update not only
but also Z, until it satisfies the inequality
|J(t + 1) J(t) < |, or the number of iterations satisfies t > Tm . At this time the clustering center Z is the result of clustering.
3
TIME ELECTRICITY PRICE
The clustering intensity is enhanced until all

the sample curves are clustered into one class,
that is, this unique clustering curve represents
the overall electric habit of the sample, and
ISBN: 978-1-941968-38-32016 SDIWC
we divides the time interval of the total sample area by this unique clustering curve. According to the load level of the unique cluster
curve, the time interval is determined by using the method of the mid-point selection.
The point on the time axis corresponding to the
middle point of the peak and valley or between
the peak and the flat is the point of the time
interval. The 24h in this area can be divided
into 3 main power use periods: General (flat)
period, peak period, and valley period.
We set the price of flat period is pf , on the basis of it, the peak and valley period under the
floating price respectively:
pp = pf (1 + )
(5)
pv = pf (1 )
(6)
Type and were floating and floating ratio.

Considering the benefits of the power supply
side and the user, we define the pull apart as
[7]:
= /, W3 /W1 1
(7)
W3 and W1 express respectively the total electricity consumption in peak and valley period. According to the implementation method
of sub-time electricity price and time division
above, we get the sub time electricity price
which is in line with the local electricity habit.
4
POWER RECOMMENDATION AND

RETURN CALCULATION
Power recommendation: Through the ideal

eliminate peak and fill in the valley, the load
curve of power system in the region will be a
parallel to the time axis line which is the curve
most wants to see. For a specific user, the actual load curve and the daily average load curve
are used. According to the comparison, it is
proposed that the use of electricity should be
less in the peak period and be more in the valley period, so that the load curve is similar to
that of the daily average load curve.
24h in a region according to the above method
is divided into n hours. The electricity price in
the N o.i period is recorded as pi . If the electricity consumption of a user in the N o.i period
41
in this area is wi , the users electricity tariff will

be:
n
X
P =
pi w i
(8)
i=1
If the user fully accept the proposal, that is, the

total electricity consumption of the user will
not change, when the peak power shifts to the
valley, the load curve will be idealized into the
daily average load curve. Note the length of the
N o.i period for ti , after accepting the proposal,
the power consumption of the N o.i period is:
n
X
Wi0 = (
wi /24) ti
5.2
Initial Position of the Natural Cluster

Centroid
We set all sampling have an ideal coverage of

all natural clusters which cover all categories.
Natural clusters that are included in the sample
have the approximate centroids like the original data set have. We Step two steps to determine the initial position of the natural cluster
centroid.
(9)
i=1
at this time the electricity becomes:

0
P =
n
X
pi wi0
(10)
i=1
The savings amount is P = P P 0 .

5
REALIZATION IN LARGE DATA ENVIRONMENT: SAMPLING DIVISION
Classical algorithms have different degrees of

restrictions on the scale of the data, and the
fuzzy curve clustering algorithm introduced in
the previous paper is no exception [8][9]. In
order to give the realization scheme of FCM
curve clustering under the big data environment, the paper proposes the idea of sampling
division based on literature [10]: using the
correlation between each sampling, not only
the independent treatment of each sample can
maintain a smaller size, but also the results can
reflect the overall processing.
5.1
Sampling
For the small data set extracted, hoping that

it can have all of the natural clusters which a
large data contained in, that is, it contains all
user types. We give the sampling formula [11]:
1
S = f n + c log( ) + c
r
1
n
1
[log( )]2 + 2 f log( ) (11)
ISBN: 978-1-941968-38-32016 SDIWC
Figure 1. Determine the initial location of the natural

cluster centroid
First step: Due to the small size of the sample,

we can use the classic algorithm in clustering
each sample, so the process of clustering FCM
curve will be very fast. Because each sampling
has the same scale, and the clustering process
is independent, so it can be implemented in
parallel processing which can reduce a lot of
running time. Set one sample of the total sample large data set (category number is c) covering c category. Because of the existence of
this inequality: 1 c0 < c, the sampling will
be clustered into c clusters, which will split a
cluster, and that is a more detailed classification situation, which will be revised in the following text.
Second step: By the first step we get c M
small clusters, and calculate the average value
42
of each small cluster:

n
i
1X
i =
ij
n j=1
6.1
(12)
ni is the data size of the cluster Ci , and ij

is the properties of a sample in Ci . Using the
mean to represent the cluster, we can get the
data set A whose size is c M . By clustering
A with improved FCM algorithm, we can get c
clusters, then it will make c M small clusters
into c large clusters. By calculating the mean
value of the c clusters, the initial position of the
natural cluster centroid can be determined.
5.3
EXAMPLE ANALYSIS
6.1.1
Demand Side Power Recommendation

Data Sources and Processing
Selecting the power consumption data with

dispersion from a county area, we use matlab
in drawing the electric power load curves from
200 users, Yuanling County, Hunan Province
(the horizontal axis is time, Y-axis is power),
as shown in Figure 2.
Means Update
Because the data is local information, the initial position of the cluster centroid often deviates from the cluster centroid of the original
data set. Therefore, it is necessary to update the
initial position by updating means. According
to the distance of the initial centroid, the remaining data samples which are not used to be
allocated to the nearest cluster:
= arg max(| i |2 )
i
(13)
are the sample properties whose class is undetermined, are the already identified categories.
When a data sample is filled with a cluster Ci
according to the above classification principle,
the mean value of the cluster is updated:
i
i ni +
ni + 1
ni ni + 1
It can be known from Figure 2, although the

individual users electricity characteristics are
not the same, but the county has a total of two
electricity peak (in the morning, afternoon), a
period of electricity valley(noon), a period of
electricity flat (night).
Therefore, we set the clustering intensity , and
the improved FCM curve clustering is used to
cluster the 200 user load curves into a class as
shown in Figure 3. The only cluster curve represents the overall power habits and levels of
the Yuanling County.
(14)
(15)
With the addition of the remaining data,

the mean position is constantly updated, and
moved to the real center of the natural cluster
until the update is complete.
At last, the data set is divided by the above
mentioned natural cluster centroid, and the
principle of dividing the data is still in the form
of the formula (13). That is, according to the
distance of the cluster centroid, we can determine the final classification.
ISBN: 978-1-941968-38-32016 SDIWC
Figure 2. Electric load curve of all users
Figure 3. After FCM curve clustering
6.1.2
Tiered Pricing for Electricity
Based on the load level (instantaneous power)

of the only cluster curve in Figure 3, and we
43
combine the reference [12][13][14] with the

midpoint method given in the previous paper to determines the time period. The result
is shown in Figure 4. The area 24h is divided
into 3 parts (5 power period): flat period (00 :
00 8 : 45) & (21 : 45 23 : 45), peak period (8 : 45 11 : 45) & (16 : 45 21 : 45),
valley period (11 : 45 16 : 45).
Figure 5. Schematic diagram of power suggestion
Figure 4. Determine time hours
Combined with the implementation measures

of Hunan power grid peak valley TOU price
and the time division above, the proposed TOU
price in line with the local electricity use is
shown in table 1.
Table 1. Time sharing electricity price
Period Type
Peak period
Flat period
Valley period
Period Type
Peak period
Flat period
Valley period
6.1.3
Range
8 : 45 11 : 45&16 : 45 21 : 45
00 : 00 8 : 45&21 : 45 23 : 45
11 : 45 16 : 45
Measures
TOU Price
Floating 80%
0.810
0.450
Down 55%
0.202
User Power Recommendation
Calculate and draw the load curve and the daily

average load curve of one user, respectively, as
shown in Figure 5 in the form of broken line
and horizontal line.
According to figure 5, the users electricity is
recommended: Reducing the power consumption according to the daily average load curve
in the period when the actual load curve is
higher than the daily average load curve. Similarly, the user ought to increase power consumption at a time below the daily average
load curve, thus participating in the demand
ISBN: 978-1-941968-38-32016 SDIWC
side management. In accordance with the previous return calculation, we can be accurately
obtained if the user followed the electricity recommendation, the ideal energy savings is 1.04
yuan per day.
The proposed power recommendation and the
amount of return, is a demand side management signal for individual users, and can be
more reasonable to strengthen the demand side
management [15][16][17].
6.2
Test under the Environment of Large

Data
In order to test the performance of sampling

partition clustering, the number of users to be
processed is increased in turn. The average
running time which processing the same user
data need are respectively tested by the sampling partition clustering and the direct clustering. We define the processing speed as v =
q/ . q is processed by the number of users.
is the average running time. Therefore, the
lifting speed ratio is = v2 /v1 .
The test and calculation results are shown in
Table 2 and figure 6.
Table 2. Example contrast test
A1
1000
2500
5000
7500
10000
13000
16000
20000
30000
B2
7.34
36.23
74.20
109.96
122.35
166.83
233.37
296.54
428.04
C3
1.44
3.36
7.50
11.08
13.98
18.57
23.19
27.84
42.94
D4
5.11
10.78
9.89
9.92
8.75
8.98
10.06
10.65
9.97
E5
136.19
69.00
67.39
68.20
81.73
77.93
68.56
67.44
70.09
F6
695.78
744.00
666.43
676.91
715.41
700.09
690.03
718.28
698.69
44
demand side response, ease the power shortage pressure, and reduce the supply and consume power cost, but also improve the systems
overall energy utilization efficiency. So it can
achieve the long-term goal of environmental
protection and sustainable development.
REFERENCES
Figure 6. Relationship between direct clustering speed,
clustering speed, lifting speed ratio and data size
From Figure 2 and table 6, we can see that the

processing speed of sampling partition clustering is always higher than that of direct clustering in the range of the data used in the experiment. In addition, when the data quantity
is higher than 10000, the lifting speed ratio
between the sampling partition clustering and
the direct clustering is stable at about 10. In
terms of large-scale data, the sampling partition clustering can still maintain a good processing speed, which proved its good processing performance.
7
CONCLUSIONS
Based on FCM clustering and sampling algorithm, the electricity recommendation model is
proposed for the demand side. Through the actual data matlab simulation test, it is proved
that in the electric power big data environment
the sampling partition clustering has a faster
processing speed. With the increase of the
amount of data, it has obvious advantages compared with the direct clustering. In addition,
the power recommendation (TOU price, electrical mode, demand side response, etc.)That
the model give is clear and accurate to meet
the users electricity habits and characteristics.
The technology can be applied to the power
demand side management in the background
of intelligent power grid with big data. It can
not only mobilize the enthusiasm of the power
1
A means Quantity
B means Average time of direct clustering
3
C means Average time of sampling partition clustering
4
D means Lifting speed ratio
5
E means Speed of direct clustering
6
F means Speed of sampling partition clustering
2
ISBN: 978-1-941968-38-32016 SDIWC
[1] Zeng Ming, Power Demand Side Management, Beijing: China Electric Power
Press, 2001:1-8.
[2] Zhang Pei, Current Situation and
Prospect of Power Big Data Application, Electrical age, 2014(12):14.
[3] Nai Jiawen, Ms D Thesis, Research
and Development of Customer Powerexpenditure Information Intelligent Analysis System, Guangdong University of
Technology, 2014.
[4] Tang Ruiwei, Application Status and
Development Prospect of Electric Power
Big Data, China Electric Power Enterprise Association Science and Technology Development Service Center:
2014:7.
[5] Yan Longchuan, Li Yaxi, Li Binchen,
Opportunities and Challenges Faced by
Power Big Data, Electric power informatization: 2013,11(4):1-4.
[6] Balazs B, Janos A, Balazs F, Fuzzy clustering and data analysis toolbox [EB],
2006[2015-10-01].
[7] Liu Xiaocong, Wang Beibei, Li Yang,
Day-ahead Generation Scheduling
Model Considering Demand Side Interaction under Smart Grid Paradigm,
Chinese Journal of Electrical Engineering: 2013,01:30-38.
[8] Li Yin. Ms D Thesis, Research and
Application of Clustering Algorithm,
Jiangnan University, 2013.
45
[9] Li Ziliu, Ms D Thesis, Research

on Real-time Data Streams Clustering Framework, Zhongshan University,
2013.
[10] Lu Zhimao, Feng Jinmei, Fang Dongmei, Novel Partition Clustering Algorithm for Large Data Processing,
Systems Engineering and Electronics,
2014,36(5):1010-1015.
[11] Sudipto G, Rajeev R, Kyuseok S, Cure:
an efficient clustering algorithm for
large databases, Information Systems,
2001,26(1):33-58.
[12] Zhu Keding, Song Yihang, Tan Fuzhong,
Design Optimization Model for Tiered
Pricing of Household Electricity Consumption, East China Electric Power,
2011,06:862-867.
[13] Li Yuan, Luo Qin, Song Yiqun, Study
on tiered level determination of TOU &
tiered pricing for residential electricity
based on demand response, Power System Protection and Control, 2012,18:6568+74.
[14] Song Yihang, Ms D Thesis, Optimization model for the design of tiered pricing for household electricity in china,
North China Electric Power University
(Beijing), 2011.
[15] Cai Pei, Weng Huiying, Yu Bin, Research on Smart Demand Side Management System in Low-Carbon Economy,
Power System Technology, 2012,10:1116.
[16] Wang Beibei, Li Yang, Demand side
management planning and implementation mechanism for smart grid,
Electric Power Automation Equipment,
2010,12:19-24.
[17] Wang Mingjun, Load Management and
Demand Side Management in Electricity Market Environment, Power System
Technology, 2005,05:1-5.
ISBN: 978-1-941968-38-32016 SDIWC
46
A Diagnosis and Prescription System to Automatically Diagnose Pests

Helin Yin 1, Da Woon Jeong2, Yeong Hyeon Gu3, Seong Joon Yoo4.*, Seog Bong Jeon5
Sejong University,
Department of Computer Engineering, Neungdong-ro, Gwangjin-gu, Seoul Korea
yinhelin0608@gmail.com1, chris410@naver.com2, yhgu@sejong.ac.kr3, sjyoo@sejong.ac.kr4.*,
sbjeon@sejong.ac.kr5
ABSTRACT
Crop losses continue to increase due to
climate change and the presence of foreign pests.
However, it is difficult for farmers to reduce
crop losses because they cannot diagnose and
prescribe against pests quickly enough.
Therefore, in order to resolve the issue, this
paper describes a mobile-based, automatic
system for pest diagnosis and prescription using
a smart device with which to diagnose pests and
obtain prescription information by taking
photographs. In order to diagnose pests, image
searches based on similarity are conducted. Due
to the features of the image-similarity search,
sufficient data sets must be obtained in order to
increase search precision. In order to increase the
pest-image data set, images of pests were
collected using a focused web crawler on the
Internet. However, because there are many
images that do not pertain to the applicable pests,
the precision of diagnosis is reduced. Therefore,
image precision was increased through an
inspection system that utilizes experts. The
images obtained with the pest-image collection
and search system is indexed using an image
similarity-based search system. Next, the
similarity of images of pests taken with the
users cell phone is compared, and information
on diagnosis and prescription is shown to the
user in real time. When the system-diagnosis
performance was measured using three crops
(pears, strawberries and grapes), a precision
level of 83% was recorded. In the future, a video
*
recognition-based system will be incorporated in

order to enhance the precision and scope of
application. It is expected that the system of
mobile-based automatic pest diagnosis and
prescription will provide quick, precise pest
information to farmers and assist in the prompt
prevention of disasters so as to minimize
economic losses.
KEYWORDS
inspection system, pest diagnosis,
based on similarity, web crawler
searches
1 INTRODUCTION
Although various smartphone applications for
the identification of crop pests are being
developed amid the rising demand for diagnosis
and prescription against loss-inducing crop pests,
such apps remain incapable of analyzing images
obtained using a smartphone in a program and
autonomously diagnosing diseases. Therefore,
in this project the images and information on
pest diagnosis and prescription were saved in a
database. Additionally, in order to collect only
the most appropriate images, unnecessary
images are eliminated through an imageinspection system in which experts participate,
and clear images of pests are selected and saved
in the "pest" database in order to provide precise
images. Through a collection system that
consumes just one-tenth of the time previously
required, the applicable image-inspection
Corresponding author
ISBN: 978-1-941968-38-32016 SDIWC
47
system increases the precision and objectivity of

pest-image data through means of an imageinspection system and the objective opinions of
semi-experts and experts based on an imageinspection system. The images transmitted by
the applicable image-inspection system are
provided to the user real-time using a similaritybased search and machine running classification
algorithm in order for the user to determine
whether the images of suspected pests are
actually so. With this system, if the user
determines that the images depict such pests, the
information on diagnosis and prescription will
be indicated by means of an app in order to
prevent any further crop losses the pests would
cause.
2 RELATED WORK
In order to conduct automatic diagnosis in an
automatic, mobile-based system for pest
diagnosis and prescription, either the method
using image similarity or the method using video
recognition could be used. For these, videos with
which to learn more about the images of pests
are necessary. In order to configure image sets
for use in study videos, it is necessary to collect
data pertaining to images of pests. Currently, in
order to configure databases of pest images,
experts must visit web pages and collect images
of pests one by one. Therefore, although the
precision level is high, the collection of such
images is time-consuming. A variety of data-including images, texts and SNS--is collected
using an open API [1]. There are many kinds of
open APIs to collect images. Among them,
Googles Open API [2] and Microsofts Open
API are most popular.
Because experts must work concurrently with
the collection of images for specific fields, there
are instances in which errors occur due to
judgments regarding ambiguous images. When
such images are used, the experts subjective
judgment is reflected, so images are not
objective. In order to resolve issues pertaining to
subjective processing methods, many methods
by which to obtain opinions objectively have
ISBN: 978-1-941968-38-32016 SDIWC
been considered in the test-processing field [3].

Among inspection systems, productinspection systems have been developed the
most, and machine vision, contents-based
image-inspection systems, etc., have been
studied [4,5]. However, there has been no
inspection system that uses or requires expert
knowledge. Therefore, it is believed that, in
certain areas, inspection systems that
incorporate expert judgment will remain
necessary.
Accordingly, for the purposes of mitigating
the shortcomings of the existing systems for
collection and inspection, systems in which
semi-experts and experts collect various images
through image-collection programs in a short
period of time and to assist semi-experts and
experts are necessary. Therefore, this paper
describes a method in which precise images are
collected with a supplementary system, pests are
diagnosed automatically and prescriptions are
provided based on an automatic mobile system.
Recently, several diagnostic systems that
employ the IT fusing technology have been
launched, and related technologies have been
proposed.
First, with [6] it is possible for a farmer to
know about the occurrence of pests on a realtime basis, and the system allows the farmer to
contrast images to conduct diagnosis or to
consult with an expert to prevent disasters.
Secondly, [7], [8] and [9] are systems that
forecast the climate, predict the occurrence of
pests, and provide pest management information.
However, with the above systems the types of
pests are limited. Moreover, the pest-diagnosis
function is not provided automatically.
Thirdly, with the research described in citation
[10], it is possible to diagnose pests in crops and
search for prescription information with a
smartphone-based pest-information search
system. However, descriptions regarding the
collection of images and recognized crops were
insufficient, and LIRE (Lucene Image Retrieval)
[11] which is similar to the image search library-was used. Consequently, the level of precision
in diagnosing pests is not high.
48
Fourthly, in the research described in citation

[12], special features of symptoms caused by
eight pests were automatically extracted, and
these features were studied and classified using
a GG vector machine (SVM), which is a
machine-learning classification algorithm, and it
is said that the average level of precision was
84%. In the research described in citation [13],
when the HOG features were studied using an
SVM classifier following an image preprocessing process, the level of precision was
97%.
Finally, in the research described in citation
[14], three cucumber diseases were classified
using a probability-based neural network model
(a type of artificial neural network), and the level
of precision was 91%.
As described, there have been many studies
on the recognition of crop pests. Although the
level of precision decreases in the vent that a
similarity-based image-search system is used,
the level of coverage is very high because many
types of pests can be handled. Although the level
of recognition precision is high for crops of
specific crops when machine learning algorithm
is used, the scope of application is narrow due to
such problems as the study time and large
amount of training data required. In order to
resolve such issues, we designed a quick but
precise mobile-based automatic pest diagnosis
and prescription system using a similarity
image-based search method and SVM, which is
a machine-learning classification method.
Figure 1. Overall system structure
3.1 Image Collection and Search-System

Design
For this paper, an Image Search API using the
Open API was used as an image-collection
program. For this paper, images were collected
using the Bing Image Search API [15] and a
database was established through use of the
image data collected through MySQL. Images
collected on the World Wide Web and the
existing images pertaining to pests were put into
a pest-image collection system. An inspection
system was used on images collected through an
image collection system in order for semiexperts and experts to determine whether the
images depict pests.
Web
Supplementary pest inspection
system
Inspection system
Semi-expert A
Web Image Crawler

Case 1
3 SYSTEM DESIGN
Semi-expert B
Case 2
Collection database
This paper describes the use of data-collection

modules to collect images of pests, similaritybased image-search modules that support fast
searches, video-recognition image processing
modules that improve precision of pest
recognition, and diagnosis and prescription
modules that provide information on diagnosis
and prescription. Figure 1 is our system structure.
ISBN: 978-1-941968-38-32016 SDIWC
Image collection system
Case 3
Similarity-based search
Expert
Image collection system
Pest-diagnosis system
Figure 2. Image-search system configuration diagram
49
The overall system configuration diagram of

the embodied pest image-search system is
shown in Figure 2. The system acts as a
supplementary system so that the pest-diagnosis
system can provide experts and semi-experts
with clear information regarding images of pests.
Table 1. Saved results, as obtained from semi-experts
and experts
The search system is saved as shown in Table

1. After semi-experts and experts in the field of
pests determine with a search system whether
the images depict pests, the results are either
saved or deleted depending on the results shown
in Table 1. In order to determine objectively and
precisely about images of pests, two semiexperts and an expert in the field of pests
determine whether the images depict pests.
Whether the images are to be saved depends on
the results shown in Table 1.
As can be seen in Table 1, cases are classified
into three types. Case 1 is the one in which two
semi-experts determine that the collected images
depict pests. In this case, because two semi-
experts distinguished clear images pertaining to

pests, the images are saved in the pest collection
database without any separate measures taken.
Case 2 is the situation in which two semi-experts
determine that it is not clear whether the
collected images depict pests. This is the one in
which the images are deleted due to the fact that
the semi-experts believe that the images are not
pertaining to pests. Case 3 is the situation in
which two semi-experts determine it is difficult
to decide whether the collected images of pests
are those of pests. Among the images, it is
difficult even for experts to distinguish images
of such diseases as pear fire blight and pear leaf
spot. Therefore, whether the images should be
saved or deleted will be determined through one
other pest experts decision for accuracy. Table
2 is an example of the tables that are saved in the
image collection system. This allows semiexperts and experts pertaining to pests to
confirm and determine images based on the
images saved in the collection system. The ID on
Table 2 is an index that indicates the number of
image data, and the types of crops number
approximately five, including pears. Diseases
were classified according to the scientific names
of pests in citation [16]. The image address
indicates the address of the images according to
the scientific names of pests, and the web
address indicates the address displayed on the
Table 2. Image Table structure
ISBN: 978-1-941968-38-32016 SDIWC
50
corresponding website. Semi-expert A and

Semi-expert B check the images of the pests
collected with a collection system in order to
determine whether the images depict pests. The
expert then checks the images in order to
determine whether the decision of the two semiexperts is correct.
similarity-based image-search system.
3.2 Design the Pest-Image Data Model

To efficiently save images of pests as well as
applicable pest information and prescription
information, the research for this paper made use
of a relational database model. Figure 3 is an ER
diagram for the saving of data. With our
proposed system, three tables were used in order
to save image information and prescription
information. The basic information and
prescription
information
regarding
the
applicable disease are produced through a query
based on the name of the disease pertaining to
the candidate pest image selected by the user.
Figure 3. ER diagram
3.3 Design of a Similarity-Based ImageSearch System

In this stage, if the user inputs images that are
believed to depict pests, similar images are
produced. In this system, LIRE (Lucene Image
Retrieval), which is an image index library will
be used to obtain image similarity.
Based on this library, the features of the
images entered are first extracted. Subsequently,
the images in the pest DB for which indexing
was completed are compared in order to produce
images with high levels of similarity. This
method is advantageous in that the speed is
conducted at a high speed, but the disadvantage
is that indexing must be conducted again when
new images are entered in the DB.
Figure 4 is a sequence diagram regarding a
ISBN: 978-1-941968-38-32016 SDIWC
Figure 4. Similarity-based search sequence diagram
The user may, with his or her smartphone,

capture images of crops that are suspected of
being infected by pests, or may select images
from a local file system. After images are
selected, the transmission button is clicked for
transmission and uploading to a web server.
Then, with the images uploaded in the server,
the degree of similarity is calculated using the
LIRE (Lucene Image Retrieval) library, and the
similarity of the many images existing in the pest
DB is calculated in order to generate six
candidate images in the order of similarity to be
shown to the user through the UI.
4 SYSTEM IMPLEMENT
In this paper, we embody a pest-image
collection-and-inspection system whereby
images are collected, identified and saved so as
to provide the applicable images for an
automatic, mobile-based pest diagnosis and
prescription system in order to select images
from the World Wide Web, particularly those
that depict suspected pests, and to produce pestdiagnosis
information
and
prescription
information.
4.1 Implement of the Pest-Image Collectionand-Inspection system
Three people two semi-experts and one expert,
determine the precision level of images through
a web page. In the event the decisions of the two
semi-experts are different, the images are
51
checked by the expert.
Figure 5. Main screen of the pest-collection/inspection

system
Figure 5 is the main screen of the inspection

system. On the main screen, those images that
are confirmed by the two experts and one expert
as depicting pests are shown on the main screen.
Based on the images displayed on the main
screen, the images that are used on mobile-based
automatic pest diagnosis and prescription
systems are obtained.
Figure 6 illustrates the work of the two semiexperts as they employ the pest-inspection
system. On that screen, the two semi-experts
determine through an inspection system whether
the images pertaining to pests could be used.
Thus the decision is based on the assessment of
the images as correct or incorrect. Additionally,
additional texts could be added for images.
Figure 6. The pest-inspection systems inspection page
Figure 7. Page on which pests are inspected again
The question of whether images can be used

is determined by the two semi-experts, as shown
in Table 1. Figure 7 is the screen on which
inspection is being conducted by the expert due
to the fact that the decisions of the two semiexperts were different. Figure 8 shows a page
that was deleted after the two semi-experts and
one expert determined that the images did not
depict pests.
Figure 8. Page being deleted by the pest-inspection
system
ISBN: 978-1-941968-38-32016 SDIWC
52
4.2 Implement of an Automatic, MobileBased Pest Diagnosis and Prescription

System
Pear fire blight
The mobile-based automatic pest diagnosis

and prescription system may be used with an app
or on the World Wide Web. Figure 9 is the initial
screen of our system. The screen is the screen UI
(User Interface) on which the user directly
selects images of suspected pests. The screen is
also the one on which types of crops are selected.
There are two methods for the selection of
images: One is to obtain images with a camera,
and the other is to obtain them from a local file
system.
After selecting images of pests through the
process shown on Figure 9, when the "Search"
button is clicked the screen (UI) as shown in
Figure 10 is displayed.
Smart Pest
Smart Pest
Choose
File
A file is not selected.
Camera
Gallery
Symptoms
Although the disease occurs on pear trees? leaves, leaf s
talks, fruits, fruit stalk, branches, and so on, usually da
mages to leaves and fruits are the greatest. At first, uncl
ear monolithic or elliptical lesions appear along the lea
ves veins, and then this symptom progresses, and count
less sooty shaped spores appear. The greatest damages t
o diseased leaves is early falling of leaves to hinder the
process of assimilation. On fruits, black spot lesions ap
pear from the early stage, and the symptom progresses
so that the lesions becomes sooty shaped. If severe, the
fruit surface becomes shaped like a scab, and the fruit b
ecomes sunken as a result, and at times deformed fruits
result.
Figure 10. Selecting images of similar pests
5 PERFORMANCE COMPARISON
To measure the performance of the pestcollection/inspection system, we compared the
number of images collected by using the
inspection system and the number of images
inspected without using the search system
during a given period of time. Additionally, the
level of precision was measured in order to
assess the performance of a similarity-based
search system on mobile-based automatic pest
diagnosis and prescription systems.
5.1 Performance of the System for Pest
Collection and Inspection
Copied to Clipboard
Figure 9. Selecting images of pests
In Figure 10, the features of the images

selected by the highest level of similarity are
produced. Among the six images of pests, the
user selects the image that is most similar in
terms of symptoms.
When the user selects an image, the pestdiagnosis
information
and
prescription
information applicable to the image are
produced.
ISBN: 978-1-941968-38-32016 SDIWC
Figure 11 shows the results of the difference

in speed when two semi-experts and one expert
searched images pertaining to pests manually
and when the search system proposed in this
paper was used to search images.
We could thus confirm that the speed at which
images were searched using the collection-andinspection system proposed in this paper was 10
times the speed at which two semi-experts and
one expert in the field of pests collected images
of pests.
53
5.2 Measurement of the Precision of the

Similarity-Based Search System
Number of images collected during the inspection time

160
152
140
In this section, the level of precision of a

similarity-based image-search system was
measured. The types of crops measured were
pears,
strawberries
and
grapes,
and
performances were measured in two cases.
Figure 12 measures the level of precision of
similarity-based search systems.
In Figure 12, there are two bar graphs for the
parameter crops. When the image with the
highest level of similarity is entered, if the
applicable image corresponds to the applicable
pest, "1" is indicated, but "0" is indicated if the
image does not correspond to the applicable pest.
"Voting" is the proportion of the applicable
images among the six that have been produced.
If the proportion is over 50%, "1" is entered, but
otherwise "0" is indicated. For example, the user
enters a query regarding an image of fire blight.
If there are five images of the fire blight out of
the six images produced, "1" is indicated
120
100
10.86X
80
60
40
14
20
0
Users searching manually
Users utilizing the search system
10 minutes of inspection time used
Figure 11. Comparison of the number of images found

when images were searched manually and when images
were searched using the search system
The speed at which images were collected was

improved through the collection-and-inspection
system proposed in this paper, and the level of
precision for the images was improved through
the opinions of the two semi-experts and one
expert in the field of pests.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Fire
blight
Black spot
Pears
Gray
mildew
Powder
mildew
Downy
mildew
Strawberries
Ranking
Phomopsis
blight
Rust
Grapes
Voting
Figure 12. Precision level of similarity-based images
ISBN: 978-1-941968-38-32016 SDIWC
54
because the number of fire blight images is

greater than three, meaning it exceeds 50% of
the images.
In this paper, performance measurements
were made for three types of crops: pears,
strawberries and grapes. In Figure 12, the
averages of the "ranking" and "voting" for each
crop were calculated. The level of precision for
"ranking" was 83%, and the level of precision
for "voting" was 78.5%.
6 CONCLUSION
Based on the above process, we designed an
embodied an automatic, mobile-based pest
diagnosis and prescription system that can be
used in crop production sites.
Images of crop pests and pest management
information were provided by the National
Institute of Horticultural and Herbal Science for
use in the proposed system. Additionally,
through the pest-image collection-andinspection system, it was possible to resolve the
issue of insufficient numbers of images to a
certain extent, and it was possible to improve the
reliability and precision of the images used for
analysis.
The proposed system made it possible to
handle more types of pests through the
similarity-based image-search system, and it
was possible to search more quickly. However,
this system did not achieve a high level of
precision in the identification of pests.
Consequently, in the future a video recognitionbased system will be incorporated for greater
precision and coverage.
Additionally, the system we propose will be
linked with the NCPMS (National Crop Pest
Management System)[17] and provide farmers
with prompt, precise pest information, thereby
assisting them with timely pest management in
order to minimize the economic losses caused by
pests.
ACKNOWLEDGEMENT
This work was supported by Korea Institute of
Planning and Evaluation for Technology in Food,
Agriculture, Forestry and Fisheries(IPET)
through (Advanced Production Technology
Development Program), funded by Ministry of
Agriculture, Food and Rural Affairs(MAFRA)
(No 315091-3) and the MSIP (Ministry of
Science, ICT and Future Planning), Korea, under
the Global IT Talent support program(IITP2016-H0905-15-1005) supervised by the IITP
(Institute for Information and Communication
Technology Promotion)
REFERENCES
[1] J.I.Kim, K.H.Lee, A Study of Contents Service
System Using Open API,International Conference
on Multimedia Information Technology and
Applications(MITA), 342-345, 2009.
[2] https://developers.google.com/products/?hl=ko.
[3] T.Brants, Inter-Annotator Agreement for a German
Newspaper Corpus, In Second International
Conference on Language Resources and Evaluation
LREC-2000.
[4] U.S.Khan, J.Iqbal, M.A.Khan, Automatic Inspection
System Using Machine Vision, 34th Applied
Imagery and Pattern Recognition Workshop
(AIPR05), pp.210-217, 2005.
[5] J.Iivarinen, J.Pakkanen, J.Rauhamaa Content-Based
Image Retrieval in Surface Inspection.
[6] http://ncpms.rda.go.kr/npms/FoodImageListR.np,
National Crop Pest Management System(NCPMS).
[7] http://nongup.gg.go.kr/works/27,Gyeonggido
Agricultural Research& Extension Service.
[8] http://www.cnnongup.net/html/kr/tech/tech_04_03.h
tml,
CHUNGCHEONGNAM-DO
Agricultural
Research & Extension Services.
[9] http://www.agri.jeju.kr/, JEJU Agricultural Research
& Extension Services.
[10] K.JAGAN, M.Balasubramanian. Recognition of
Paddy Plant Diseases Based on Histogram Oriented
Gradient Features. International Journal of Advanced
ISBN: 978-1-941968-38-32016 SDIWC
55
Research in Computer and Communication

Engineering, 2016; 5(3), pp.1071-1074.
[11] S.Yun, X.F.Wang, S.W.Zhang, C.L.Zhang. PNN
based crop disease recognition with leaf image
features and meteorological data. Int J Agric & Biol
Eng, 2015: 8(4), pp.60-68.
[12] Jagadeesh D, Rajesh and Abdulmunaf S.
Classification of Fungal Disease Symptoms affected
on Cereals using Color Texture Features.
International Journal of Signal Processing, 2013; 6(6):
pp.321-330.
[13] J.H.Kang, S.H.Jung, S.S.Nor, W.H.So, C.B.Sim.
Design and Implementation of Produce Farming
Field-Oriented Smart Pest Information Retrieval
System based on Mobile for u-Farm. 2015, pp.11451156.
[14] M.Lux and Savvas A. LIRe: Lucene Image Retrieval
- An Extensible Java CBIR Library. Proceeding MM
'08 Proceedings of the 16th ACM international
conference on Multimedia. 2008, pp.1085-1088.
[15] https://www.bing.com/dev.
[16] https://ko.wikipedia.org/wiki/ Scientific.
[17] http://www.nihhs.go.kr/english/index.asp, National
Institute of Horticultural and Herbal Science.
ISBN: 978-1-941968-38-32016 SDIWC
56
Monitoring of events in SOA architecture for Real Time Financial Decision System
Dominik Ryko* and Pawe Gawiski**
*Warsaw University of Technology, **Softman
*Nowowiejska 15/19, 00-665 Warsaw, **Gen. Okulickiego 5F, 05-500 Piaseczno
*d.ryzko@ii.pw.edu.pl, **pawel.glawinski@softman.pl
ABSTRACT
Most of the financial decision systems implemented
to day are based on aggregation of transactions in
data warehouses or other reporting environments and
then generating appropriate reports for the analysts
and managers. This makes the decision cycle, as
counted from an individual transaction or event, very
long. On the other hand modifying functional
workflows to embed financial controlling steps
imposes technological challenges related to change
management and additional delays. The paper
proposes an innovative approach to solving the
above problems, by implementing a multi-agent
system for monitoring of financial systems and
providing real-time, flexible and efficient way for
decisions. The research was conducted by Warsaw
University of Technology and Softman as a part of a
project financed by The National Center for Research
and Development in Innovative Economy
Programme (POIG) measure 1.4 project No:
POIG.01.04.00-14-061/12.
KEYWORDS
Service-Oriented
Architecture,
Mutli-Agent
Systems, distributed systems, real time analytics,
controlling systems
1 INTRODUCTION
In this Chapter motivation for the research as
well as previous approaches found in the
literature are described. The business need
which is solved by the proposed solution is also
analysed.
Service-Oriented Architecture (SOA) is an

architectural style that supports serviceorientation.
Service-orientation is a way of thinking in terms
of services and service-based development and
the outcomes of services.
A service:
Is a logical representation of a
repeatable business activity that has a
specified outcome (e.g., check customer
credit, provide weather data, consolidate
drilling reports)
Is self-contained
May be composed of other services
Is a black box to consumers of the service
The migration of architectures from monolith to
Service Oriented Architectures (SOA) resulted
in the change of the way data is exchanged in the
systems. Rather than through shared databases,
services communicate with the use of well
defined APIs [3]. In the full SOA
implementation a Enterprise Service Bus (ESB)
plays the role of the component responsible for
delivery of data according to the defined
business processes. The following specific tasks
can be handled by the ESB [4]:
Providing connectivity
Data transformation
(Intelligent) routing
Dealing with security
Dealing with reliability
Service management
Monitoring and logging
1.1 Motivation
The Open Group defines formally SOA in the
following way:
ISBN: 978-1-941968-38-32016 SDIWC
57
services into the Raw Data Store. Then it is

processed by the transformation backend and
put into the reporting store, usually containing
joined and aggregated data. Finally, an SQL
output endpoint is provided in order to plug in
ad-hoc SQL and reporting tools.
Figure 1
Performing advanced, real time analysis in such
a setting is a challenging task. Traditionally this
can be done by designing analytical or
controlling steps in the business process
definition and building specialized services to
perform these tasks. However, this approach has
several downsides. Firstly, any change in the
analytics changes the process and therefore
requires retesting to avoid introduction of bugs.
Secondly, there is an overhead which affects the
performance of the tasks which need to be
performed. Finally, it limits the possibility to
define analytical rules in a more generic,
descriptive way, understandable to nontechnical people.
Crucial for the ability to perform the analytical
tasks, is proper interpretation of the data. It is not
sufficient to capture all the messages exchanged
in the system, but they have to be interpreted
according to their context and content.
Therefore, the issue of semantics is fundamental
to this work and is further elaborated.
The motivation for this work is to analyse the
possibilities to apply Multi-Agent System
paradigms to build analytical functionalities as
an additional layer of the system, which can
monitor the business processes by plugging into
the service bus and intercepting the data from
the messages exchanged over it.
1.2 Previous Work
In [5] Aggregated Reporting pattern for SOA is
described. The pattern is designed to overcome
the distribution of data across services by
creating a service that gathers immutable copies
of data from multiple services for reporting
purposes. Figure 2 shows the pattern of the
architecture. The service works as follows.
Firstly, the data is transferred from the source
ISBN: 978-1-941968-38-32016 SDIWC
Figure 2
Four different ways of getting the data into the
Aggregated Reporting are proposed:
Actively calling other services - use of
other services contracts to get new data
Passively getting data from services subscribing to batch data exports or
events
Service SQL push - services export a
view of internal data
ETL SQL push - as in the option above
but with the involvement of external
ETL tool
The advantage of using Aggregated Reporting
include:
Holding of immutable data with
possible versioning if changes are
received
Providing single SQL interface for reporting
Possibility to highly optimize reporting

efficiency
As for the disadvantages:
High complexity of the solution
Relatively large latency in data access
Duplication of data
In [8] Schiefer and Seufert propose sense &
respond loops which enables real-time analytics
across corporate business processes. Five
58
distinct stages of data processing are proposed in

this approach:
Sense - Which is the current state of the

business environment?
Interpret - What do the captured data
indicate?
Analyse - Which business opportunities
and risks can arise?
Decide - Which strategy is the best to
improve the current situation of the
organisation?
Respond - Who has to implement the
decision?
As opposed to this work, the sense & response

functionalities are implemented as additional
services, which communicate with each other
over the ESB. The work leaves out the
technological details of the implementation, so
it is difficult to asses how it can be applied to
systems provided by various vendors.
The use of Multi-Agent systems for monitoring

in SOA has been analysed before. Monitoring of
performance and SLA was analyzed in [2].
Agents are grouped in clusters assigned to
particular Web Services and act as proxies
between WS and the client. The approach allows
redirection of requests in case of problems, but
creates significant overhead and, by becoming
part of the flow, agents can generate problems
by themselves.
In [1] Intelligent agents located across the
system perform asynchronous, distributed
measurements separately for each selected
service or process. Compliance with required
performance measures or SLAs can be defined
and appropriate warnings are generated and
reported. The approach uses Aspect-Oriented
Programming to plug into the application server
and overlook the flow of processes in the system.
ISBN: 978-1-941968-38-32016 SDIWC
This work extends the above research by

plugging into the service bus via documented
APIs, which gives a reliable and sustainable
model for integration. Also the monitoring goes
much further than in the case of performance
monitoring. Here, the actual data has to be
accessed, interpreted and processed.
2 Business Need
The risk is a basis of running a business and a
condition of market mechanisms working
properly. In terms of the business operating on
the financial market the risk should be managed.
The lack of proper mechanisms defined as risk
management processes can lead businesses to
substantial losses or even a bankruptcy.
Spectacular examples of the lack of the efficient
supervision of the risk [6] are connected with
using advanced financial instruments. As a
consequence, it led to insolvency such
businesses like Barrings Bank (the oldest
English commercial bank) or LCTM Fund. Less
spectacular examples resulting from operational
errors or frauds are often concealed by
institutions.
The exposure of the business to the risk on the
financial market depends on many factors which
can be defined by the business profile. Based on
research done recently [6], it can be concluded
that the operational risk is more and more
significant in the banking activities. It is visibly
less significant than the credit risk, but its greater
importance in the risk portfolio is connected
with both the improvement in the quality of
credit risk management methods and changes in
business model. The distribution of the main
risks in the banking activities based on recent
research looks as follows:
credit risk - 60%
operational risk - 25%
market risk - 15%
Operational risk will be defined (according to
[7]) as the loss risk incurred from internal
processes functioning improperly, people,
systems and external events. The control of this
59
risk is directly connected with business model

used by this business.
Lets assume the process of granting financial
instruments is conducted on a 24/7 basis with the
instant time of decision and immediate
disbursement and the business process itself is
highly or fully automated. Then, it requires the
control and the risk management to be conducted
in the automated way and it concerns both the
transaction and the appropriately defined
portfolio of these transactions.
The crucial elements connected with the risk
management of the above example of granting
the credit instrument are integrated directly with
a fundamental business process supported by the
IT system. The list of crucial elements includes:
Verification of the customer in debtor
databases, credit information bureaus, as
well as
blocked identification
documents bases. These verification is
performed in the context of managing the
operational risk.
Scoring performed by the IT system for
the purpose of the credit risk
management.
An automatic model of granting a credit, the use
of services provided by different external
suppliers in the business process (e.g. based on
SOA architecture) and most importantly, the use
of Internet channels of communication with the
customer make this particular business model
more vulnerable to frauds and software errors.
If we implement the above described business
process based on the chosen engine of the
business processes and we integrate all steps of
the process based on SOA architecture, we
should supplement this model with control
mechanisms over the implementation of
particular business processes instances as well as
sets of these instances. Control mechanisms
should enable changes of particular procedures
to improve them and create new forms of control
without changes in the basic business process.
Control mechanisms should be based on
information provided dynamically as part of the
communication in the business process.
ISBN: 978-1-941968-38-32016 SDIWC
If we assume that complete information in the

process consists of e-documents and the process
messages then the control mechanism
implementation model can be based on the
analysis of these forms of information.
The suggested solution is based on capturing
and analysing information of the business
process in the context of this process instance
and the defined time window.
3 SYSTEM ARCHITECTURE
3.1 Data Interception
Possible technologies to capture the data
communication over ESB include:
Use fo APIs
Proxy
Aspect Oriented Programming
In the project feasibility of applying these
techniques for various environments was tested.
Namely Oracle, IBM and JBoss service buss
were analysed. The conclusions show that there
is no single mechanism, which can be used in
each scenario.
APIs provide the most controlled environment
and are supported by the supplier of the
technology, but are usually limited with respect
to the functionality and the scope of data. There
no simple way to fix this.
Proxy allows the most generic approach, in
which we capture all communication and
process it, extracting only the data which is
needed. The downside, is that we are prone to the
changes in protocols and data formats,
something which can be transparent if we use
APIs. A single message can be also stripped of
some context, which is provided to the service
by another mechanism, thus semantics can
become ambiguous.
Aspect Oriented Programming (AOP) is
probably the most advanced of the methods. It
uses Load-Time Weaving to execute a code at
join points (eg. method execution) [7]. It allows
to introduce some computation at the locations
where it was not prepared by the vendor, so e.g.
60
can reach the data not provided through the API.

The major challenge is finding the right join
points and also to avoid introduction of new
bugs to the system.
Since the analysis showed no single good way of
extracting the data, we propose to create
dedicated plugins for particular products, which
can use any of the technologies described above.
The plugins should be able only to extract the
data and transform it into a common format. This
allows us to abstract from the particular
technology and build further common
components of the system.
E.g. for Oracle OSB - Oracle Web Service
Manager (OWSM) policies have been chosen as
the best mechanism. OWSM provides
capabilities to build, enforce, run and monitor
Web service policies, such as security, reliable
messaging, MTOM, and addressing policies. It
is possible to declaratively attach policies to
particular WSs. They also provide wide
possibilities for defining parameters (constants,
global and service level).
For JBoss, Message Store or Pipeline
interceptors can be used. The first one seems to
be more efficient, but is more memory heavy
since entire messages are serialized. In the case
of Pipeline Interceptors, it is possible to select
specific fields to be captured and therefore to
save on the data size.
IBM Integration Bus provides two native
mechanisms for message monitoring and
capturing. It is also possible to modify the
message flow and redirect a message to a
different application.
It is also possible to perform data capture and
analysis in case of a loosely-coupled service
architecture without the ESB. In such a case
services communicate directly with each other
when needed. In this case there is no ready
mechanism to do so. Therefore, using a proxy to
capture
the
messages
is
possible.
The most important property of all the approach
to produce dedicated plugins using the
mechanisms described above, is that as an output
of each of them we can define a common XML
format with specified data fields, which will be
ISBN: 978-1-941968-38-32016 SDIWC
processed in the further analytics. Therefore a

common administration mechanism has been
created, which allows specification of the
message fields and metadata, which should be
captured by the plugins.
3.2. Overall architecture
The overall architecture of the data extraction
system is depicted in Figure below.
Figure 3
Data can be captured from various systems by
the plugins according to the predefined
configuration. Then it is physically extracted and
enriched with metadata, such as time, subsystem
etc. Such prepared data is transferred to the
analytical modules.
3.3 System performance
Performance requirements for the system were
estimated based on the analysis of business
processes on the financial services market. The
maximum number of financial instruments of
the credit type granted by leading companies on
the financial market doesn't exceed one
thousand. Conclusion of such a number of
agreements depending on the availability and
attraction of the product can be associated with
the need for substantial amount of operational
processes. We have assumed that 10 000
processes of granting the credit instrument will
be initiated. The process consists of around 25
steps connected to web services. A scope of
business and operating information is on average
50 elements. For the daily after-sales service we
should adopt the same information requirements
as for sale processes. To sum up, we create one
61
million of XML messages daily, each of the

average content of 50 information elements with
simple data (like number, date, content). The
system was design to handle such load and to
scale up for potentially more complex scenarios.
4. CONCLUSIONS
The paper describes a system for monitoring of
events in SOA architecture for Real Time
Financial Decision System. Means for
intercepting of the data have been analyzed and
a plugin layer was introduced to manage this
task. The captured data is enriched with
metadata and passed to a multi-agent system for
further analysis.
The overall approach is flexible and robust,
imposes no computational overhead and its
technology dependence is reduced to minimum.
It allows separation of controlling steps from the
core business processes, which results in
minimizing the re-testing need and therefore
simplifies the change management.
4. N. M. Josuttis. SOA in practice: the art

of distributed system design. O'Reilly
Media, Inc., 2007.
5. A. Rotem-Gal-Oz, E. Bruno, and U.
Dahan. Co, SOA Patterns. Manning
Publications 2012
6. D. Gtarek, R. Maksymiuk, M. Krysiak,
. Witkowski, Nowoczesne metody
zarzdzania ryzykiem finansowym,
WIG-Press, Warszawa 2001
7. K. Jajuga, Zarzdzanie ryzykiem, PWN
Warszawa 2009
8. J. Schiefer and A. Seufert, Management
and Controlling of Time-Sensitive
Business Processes with Sense &
Respond. In International Conference on
Computational
Intelligence
for
Modelling, Control and Automation and
International Conference on Intelligent
Agents, Web Technologies and Internet
Commerce (CIMCA-IAWTIC'06) (Vol.
1, pp. 77-82). IEEE, (2005, November).
REFERENCES
1. D.Ryko and A.Ihnatowicz. Multi-agent
Approach to Monitoring of Systems in
SOA Architecture. In New Challenges
for Intelligent Information and Database
Systems (pp. 309-318). Springer Berlin
Heidelberg, 2011
2. A.Miede,
N.Repp,
J.Eckert,
R.Steinmetz, Cooperation Mechanisms
for Minitoring Agents in Serviceoriented Architectures. Tagungsband der
9.
internationalen
Tagung
Wirtschaftsinformatik 2009, vol. 1, p.
749-758,
sterreichische
Computer
Gesellschaft, February 2009
3. M. P. Papazoglou. Service-oriented
computing: Concepts, characteristics and
directions, Web Information Systems
Engineering, 2003. WISE 2003.
Proceedings of the Fourth International
Conference on , pages 312. IEEE, 2003.
ISBN: 978-1-941968-38-32016 SDIWC
62
A Study of Data Acquisition and Analysis for Drivers Behavior and

Characteristics through Application of Smart Devices and Data Mining
Hideto Kubota1
1
Cmicro Inc.
Masakazu Kyokane2 Yoshiro Imai3 Kazuaki Ando3 Shinichi Masuda1

2
3
Fujitsu Info Tech Inc.
Kagawa University
3
2217-20 Hayashi-cho, Takamatsu city, Kagawa pref. 761-0396 Japan
E-mail(corresponding authors 3 ): {imai}@eng.kagawa-u.ac.jp
ABSTRACT
Recently, several kinds of smart devices are coming available at almost every environment. And
Data mining becomes very much popular and useful for behavior analysis and decision making. In
this study, we will apply some smart devices, data
mining approach and visualization into treatment of
driving characteristics. Data acquisition of driving
characteristics has been achieved by means of smart
devices, such as iPhone and Android device. And
its data analysis has been executed together with
simple SQL database and some kind of statistical
data processing facilities. Now we will try to visualize driving characteristics and extract some useful
information for drivers/ road managers through the
above data acquisition and analysis.
KEYWORDS
Smart Phone, Visualization of Behavior, Machine
Learning.
INTRODUCTION
Day by day, the presence of the car has become indispensable as shifting means of our
everyday life. Although traffic accident occurrence number tends to decrease, but 40% of
such accidents have occurred at the traffic intersections. Almost drivers have to be careful
for such intersections because they are the very
places where several kinds of drivers skill encounter one another nearly at the same time.
Drivers from beginners to proficiency must
get their hands around driving habits, tendencies and/or characteristics in order to avoid
improperly-driving behavior, which may probably lead to traffic accidents. So they will need
to visualize their driving behavior and characteristics as easily as possible. As you know,
ISBN: 978-1-941968-38-32016 SDIWC
to use smart devices will be one of the most

promising candidates to acquire characteristic
data at the driving simply, efficiently and effectively.
A very famous article[1] presents Mobile
phones or smartphones are rapidly becoming
the central computer and communication device in peoples lives. Importantly, todays
smartphones are programmable and come with
a growing set of cheap powerful embedded
sensors, such as an accelerometer, digital compass, gyroscope, GPS, microphone, and so on.
And not only its authors but also almost all
of us can believe that sensor-equipped mobile
phones will revolutionize many sectors of our
economy, including business, healthcare, social networks, environmental monitoring, and
transportation.
In this study, we have applied some smart devices such as Apple iPhone and Android-based
smart phone to sensing or monitoring many
kinds of data for driving characteristics, employed data mining approach for them in order
to analyze drivers behavior, and visualize such
a behavior by means of machine learning such
as k-Nearest Neighbors algorithm and Support
Vector Machine.
This paper describes our study about data acquisition from smart devices, analysis of data
for drivers behavior or driving characteristics
through machine learning. The next section
describes related works about application of
smart phone-based data acquisition and analysis. The third one illustrates system configuration for our study. The fourth one introduces practically acquired data from our experiments. The fifth one demonstrates analysis and consideration of such acquired data by
means of machine learning. And finally the last
one includes our summaries.
63
RELATED WORKS
Aude Hofleitner, Ryan Herring, Alexandre

Bayen of UC Berkeley [2] presented a hybrid
modeling framework for estimating and predicting arterial traffic conditions using streaming GPS probe data. Their model was based
on a well-established theory of traffic flow
through signalized intersections and was combined with a machine learning framework to
both learn static parameters of the roadways
(such as free flow velocity or traffic signal parameters) as well as to estimate and predict
travel times through the arterial network. The
machine learning component of the approach
used the significant amount of historical data
collected by the Mobile Millennium system
since March 2009 with over 500 probe vehicles reporting their position once per minute in
San Francisco, CA. Their hybrid model provided a distinct advantage over pure statistical
or pure traffic theory models in that it was robust to noisy data (due to the large volumes of
historical data) and it produced forecasts using
traffic flow theory principles consistent with
the physics of traffic. Validation of the model
was performed in two different ways. (1) a
large scale test of the model was performed
by splitting the data source into two sets, using the first to produce the estimates and the
second to validate them. (2) an alternate validation approach was presented. It consisted
of a 3-day experiment in which GPS data had
been collected once per second from 20 drivers
on four routes through San Francisco, allowing
for precise calculation of actual travel times.
The model was run by down-sampling the data
and validated using the travel times from these
20 drivers. The results indicated that this approach was a significant step forward in estimating traffic states throughout the arterial network using a relatively small amount of realtime data. The estimates from their model were
compared to those given by a data-driven baseline algorithm, for which they achieved a 16%
improvement in terms of the root mean squared
error of travel time estimates. The primary
reason for success was the reliance on a flow
model of traffic, which ensured that estimates
ISBN: 978-1-941968-38-32016 SDIWC
were consistent with the physics of traffic.

Mohamed Fazeen of the University of North
Texas and his colleagues [3] described,
As vehicle manufacturers continue to increase
their emphasis on safety with advanced driverassistance systems (ADASs), we propose a device that is not only already in abundance but
portable enough as well to be one of the most
effective multipurpose devices that are able to
analyze and advise on safety conditions. Mobile smartphones today are equipped with numerous sensors that can help to aid in safety
enhancements for drivers on the road. In this
paper, we use the three-axis accelerometer of
an Android-based smartphone to record and
analyze various driver behaviors and external
road conditions that could potentially be hazardous to the health of the driver, the neighboring public, and the automobile. Effective use
of these data can educate a potentially dangerous driver on how to safely and efficiently operate a vehicle. With real-time analysis and auditory alerts of these factors, we can increase a
drivers overall awareness to maximize safety.
Jin-Hyuk Hong, Ben Margines, Anind K. Dey
of Carnegie Mellon University [4] reported,
In order to understand and model aggressive
driving style, we construct an in-vehicle sensing platform that uses a smartphone instead of
using heavyweight, expensive systems. Utilizing additional cheap sensors, our sensing
platform can collect useful information about
vehicle movement, maneuvering and steering
wheel movement. We use this data and apply machine learning to build a driver model
that evaluates drivers driving styles based on
a number of driving-related features. From
a naturalistic data collection from 22 drivers
for 3 weeks, we analyzed the characteristics of
drivers who have an aggressive driving style.
Our model classified those drivers with an
accuracy of 90.5% (violation-class) and 81%
(questionnaire-class).
Erez Shmueli and his colleagues of
MIT(Media Lab.)
[5] described, The
ability to understand social systems through
the aid of computational tools is central to
the emerging field of computational social
64
systems. Such understanding can answer

epistemological questions on human behavior
in a data-driven manner, and provide prescriptive guidelines for persuading humans to
undertake certain actions in real-world social
scenarios. The growing number of works
in this subfield has the potential to impact
multiple walks of human life including health,
wellness, productivity, mobility, transportation, education, shopping, and sustenance.
They also gave us contribution of their paper,
which was twofold, namely, (first) providing
a functional survey of recent advances in
sensing, understanding, and shaping human
behavior, focusing on real-world behavior of
users as measured using passive sensors and
(second) presenting a case study on how trust,
which was an important building block of
computational social systems, could be quantified, sensed, and applied to shape human
behavior. Their findings suggested that:1) trust
could be operationalized and predicted via
computational methods (passive sensing and
network analysis) and 2) trust had a significant
impact on social persuasion; in fact, it had
been found to be significantly more effective
than the closeness of ties in determining the
amount of behavior change.
3
3.1
CONFIGURATION,
MANIPULATION, AND DATA ACQUISITION
system configuration and dataflow
As known in the previous sections, smart devices can allow us to acquire the following
sensored data such as GPS, compass, gyroscope, accelerometer, and so on. For example, data from GPS facility can provide us with
not only navigation but also position-located
several useful information. And information
from compass, gyroscope and accelerometer
can bring the other following location-based
and/or movement-sensing applications to us,
such as detecting direction and/or orientation,
characterizing behaviour and habit of movements, and recognizing physical situation and
environment.
So we have begun to acquire several kinds
ISBN: 978-1-941968-38-32016 SDIWC
Figure 1. System Configuration about Data Acquisition

between Smart device and Web server.
of data and information from smart devices

by means of our original software application, transfer such data to the specified server
through HTTP connectivity between smart device(s) and server, and cumulate them into
our special Web server based on simple SQL
database scheme. Figure1 shows our system
configuration about data acquisition between
smart device and Web server. At the server, we
can perform data mining by means of retrieving and extracting specified data from a series
of data stream from the beginning to the end
stored SQL database.
3.2 manipulation with sliding windows
Well-known static sliding window approach
uses fixed-length temporal areas called windows that shift to focus each window from left
to right sequentially. Operation of shift is to
move focused window, so it used to be called
sliding window. Each window position produces one area, namely limited segment, that
is used to isolate/extract a block of data, or a
series of records.It uses two parameters: the
windows length(=wl) and the shift step(=ss).
Figure2 shows our (static) sliding window approach to extract each block of data in order to
detect specific phenomenon or recognize characteristic behavior from left to right sequentially. Our approach defines 32 sampling size
of data block for wl and half size of window,
namely 16 size, for ss. We have employed
half-size overlapping windows for sliding window approach because it is easy and efficient
for us to find specific phenomenon in order to
avoid misjudgment at the boundary point between the previous and the next windows, al65
Figure 2. Employment of sliding window approach with

half-size overlapping to extract focused data block and
to recognize characteristic behaviour.
though it may pay approximately double computing costs.

We have plaid our attention to driving behaviors at the traffic intersections because many
accidents or blood-freeze-feelings frequently
occur at that places. Normally, drivers have decided to go straight through, make a right turn,
or make a left turn at the traffic intersection. In
this study, we have focused on driving to make
a left turn at the intersection. So, we must detect or recognize driving behavior of making a
left turn automatically through scanning of acquired data files from smart devices.
With sliding window approach, we can recognize driving behavior of making a left turn
with more precision as well as efficiency. Figure 3 shows how to recognize making a left
turn with comparison between left and right
halves in the Time Window using average A1
and A2 of azimuth direction values for left half
and right one respectively.

We have defined some suitable value for
threshold T r. And we will be able to obtain
computed result of recognition for a left turn at
the case where A2 A1 < T r or a right turn
where A2A1 > T r. So we can cast anchor at
the suitable locations on the series of acquired
data set just like specifying an effective timestamp for each point when driver made a left
turn for acquired data from smart device. After that, we can utilize some anchored data set
and perform data mining approach for recognize driving behavior in order to classify what
a kind of making a left turn based on anchored
position of several mounts of sensored data or
information. The next section illustrates four
kinds of patterns of making a left turn at the
intersection and practically acquired data for
making a left turn.
3.3 data acquisition from experiment
We have defined the following four patterns of
making a left turn at the intersection.
Pattern No1 of Left turn: making a left
turn at the traffic intersection with NO acceleration(Abbreviation: LtNo). Figure
4 shows the pattern No1. Green-fill box
specifies a starting position for inbound
into left turn and green-frame box specifies an ending position for outbound from
left turn in Figure 4 7.
turn at the traffic intersection with ACceleration (Abbreviation: LtAc). Figure 5
shows the pattern No2.
Figure 3. Comparison between left and right halves in

the Time Window Using each average A1 and A2 of azimuth direction value(s).
ISBN: 978-1-941968-38-32016 SDIWC
Figure 4. LtNo
Figure 5. LtAc
66
Pattern No3 of Left turn: making a

left turn at the traffic intersection with
BRakeage (Abbreviation: LtBr). Figure
6 shows the pattern No3.
turn at the traffic intersection with OVerrunning (Abbreviation: LtOv). Figure 7
shows the pattern No4.
Table 1. Detail of Acquired Data

experiment
numbers of experiments
acquired left-turn data file
defective data
unadapted data
mismatch driving data
usable data
acquired data
18 times
72 (CSV files)
4 sets
4 sets
6 sets
58 sets
6 sets of mismatch driving data in such 72 sets

of files. So we must apply our data mining approach to the remaining 58 sets of data files and
analyze them in order to recognize the relevant
driving behavior about making a left turn at the
intersection.
3.4 parameters for pattern recognition
Figure 6. LtBr
Figure 7. LtOv
Figure 8 shows our experimental driving

course, and we can have four times of necessary data sets from one-time drive on the circuit. It takes about 15 minutes to drive our
experimental car along the yellow-lined square
circuit. We used to drive our car at the speed
approximately 40-50 Km/hour, which is a natural speed in the city.
At first, we did not know which parameters

were necessary or sufficient for our application
of pattern recognition about driving behaviour.
1. Average of movement intensity: accelerations for x-, y-, and z-axis; ones for pitch,
roll and yaw. (3+3 = 6 elements)
2. SD: standard deviations for x-, y-, and zaxis; ones for pitch, roll and yaw. (3+3
= 6 elements)
3. Maximal power spectra for x-, y-, and zaxis; ones for pitch, roll and yaw. (3+3
= 6 elements)
4. Kinetic energy for x-, y-, and z-axis; ones
for pitch, roll and yaw. (3+3 = 6 elements)
Figure 8. our experimental driving course.
From practical driving experiments, we have

acquired several numbers of data described in
Table 1. The number of experiments were 18
times, and we had four sets of data about making a left turn at the intersection from one time
experiment. Table 1 includes 72 sets of acquired left-turn data files, but there were 4 sets
of defective data, 6 sets of unadapted data, and
ISBN: 978-1-941968-38-32016 SDIWC
5. RMS: root-mean-square for x-, y-, and zaxis; ones for pitch, roll and yaw. (3+3
= 6 elements)
6. Average intensity of 3-axis composite
value and pitch/roll/yaw composite value.
(1+1= 2 elements)
7. Dispersion intensity of 3-axis composite
value and pitch/roll/yaw composite value.
8. SMA: Normalized Signal Magnitude
Area[6][7] of 3-axis composite value and
67
pitch/roll/yaw composite value. (1+1=

2 elements)
So we had begun to compute 36 dimensional
elements of data set for initial pattern recognition for classification of driving behavior. The
next section demonstrates practical computation of machine learning about such remaining data in order to perform pattern recognition
and visualize tendency of the relevant driving
behavior.
4
ways to divide the original sample into a training and a validation set. And we had obtained
good recognition rate(R = 52.73%) with empirically optimal k = 7 after choosing k to
be an odd number from 1 to 19 via bootstrap(=counting up) method. Table 2 is a result of average of rate for right answers from
8-fold cross-validation after every single subset is retained as the validation data for testing
the model, and other remaining 7 subsets are
used as training data consequently.
Table 2. Classification of Driving Behaiviour
ANALYSIS AND CONSIDERATION
4.1
classification by machine learning(I)
We have employed the k-Nearest Neighbors

algorithm (or k-NN for short) and Support
Vector Machines (SVM for short) for pattern
recognition or classification of driving behavior among the following four patterns: LtNO,
LtAc, LtBr and LtOv, which are introduced in
the previous section. The former is a wellknown non-parametric method used for classification in pattern recognition and the input
consists of the (given value: ) k closest training
examples in the feature space. k-NN is a type
of instance-based learning, or lazy learning,
where the function is only approximated locally and all computation is deferred until classification. The k-NN algorithm is among the
simplest of all machine learning algorithms.
The latter is another well-known supervised
learning model with associated learning algorithms. It can analyze data used for classification analysis1 in machine learning.
In k-NN classification, the output is a class
membership. An object is classified by a majority vote of its neighbors, with the object being assigned to the class most common among
its k nearest neighbors (k is a positive, typically
small integer.).
We had employed exhaustive cross-validation
methods which learn and test on all possible
1
Given a set of training examples, each marked for

belonging to one of two categories, an SVM training algorithm builds a model that assigns new examples into
one category or the other, making it a non-probabilistic
binary linear classifier.
ISBN: 978-1-941968-38-32016 SDIWC
LtNo
LtAc
LtBr
LtOv
Driving Characteristics by k-NN

LtNo
LtAc
LtBr
LtOv
43(48.9%) 13(17.6%) 13(18.6%) 10(12.7%)
21(23.9%) 36(48.6%) 8(11.4%) 14(17.7%)
12(13.6%) 14(18.9%) 39(55.7%) 10(12.7%)
12(13.6%) 11(14.9%) 10(14.3%) 45(57.0%)
An SVM model is a representation of the examples as points in space, mapped so that the
examples of the separate categories are divided
by a clear gap that is as wide as possible. New
examples are then mapped into that same space
and predicted to belong to a category based on
which side of the gap they fall on.
We had employed SVM with Gaussian kernel applied to the usable dataset given in Table 1. And we had obtained good recognition
rate(R = 55.94%) with gamma value: =
0.05011872 and cost value: C = 50.11872.
Table 3 is a result of average of rate for right
answers based on the condition with the above
good recognition rate(R = 55.94%).
Table 3. Classification of Driving Behaiviour
LtNo
LtAc
LtBr
LtOv
Driving Characteristics by SVM

LtNo
LtAc
LtBr
LtOv
76(96.2%)
2(2.9%)
5(6.7%)
2(2.6%)
1(1.3%)
73(92.4%)
2(2.7%)
1(1.3%)
1(1.3%)
1(1.3%)
62(82.7%)
4(5.1%)
1(1.3%)
3(3.8%)
6(8.0%)
71(91.0%)
4.2 dimensionality reduction based on

principal component analysis
Another effect of high dimensionality on distance functions concerns k-NN graphs con68
structed from a data set using a distance function. As the dimension increases, the degree
of distribution of the k-NN digraph becomes
skewed with a peak on the right because of
the emergence of a disproportionate number of
hubs, that is, data-points that appear in many
more k-NN lists of other data-points than the
average. This phenomenon2 can have a considerable impact on various techniques for classification (including the k-NN classifier), semisupervised learning, and clustering, and it also
affects information retrieval.
Principal component analysis (PCA) is a statistical procedure that uses an orthogonal transformation to convert a set of observations of
possibly correlated variables into a set of values of linearly uncorrelated variables called
principal components. Our PCA analysis has
revealed that the top 5 components explains
approximately 87.76% of the usable data set
(shown in Table 4). It is sure that reduction
of the total dimension can be carried our by
elimination of non-contributed component(s)
according to the above PCA, because Table 4
shows analyzed results by our PCA and indicates that the score of Cumulative Proportion(CP) is more than 87% within the 5 components,
1. Average of movement intensity: accelerations for x-, y-, and z-axis; ones for pitch,
roll and yaw. (3+3 = 6 elements)
2. Maximal power spectra for x-, y-, and zaxis; ones for pitch and roll. (3+2 = 5
elements)
3. RMS: root-mean-square for y- and z-axis;
one for roll. (2+1 = 3 elements)
Consequently, we are now ready to compute
the above 14 dimensional elements of data set
for improved pattern recognition for classification of driving behavior.
4.3 classification by machine learning(II)
At k-NN application to recognition of driving
characteristics, at first, we have obtained good
recognition rate(R = 58.16%) with empirically optimal k = 3 after choosing k to be an
odd number from 1 to 19 via bootstrap method
and Table 5, which is a result of average of
rate for right answers from exhaustive crossvalidation after every single subset is retained
as the validation data for testing the model, and
other remaining subsets are used as training
data consequently.
Table 5. Improved Classification of Driving Behaiviour
Table 4. Standard Deviation, Proportion of Variance,

Cumulative Proportion of the top 5 components of the
usable data set in Table1
the top 5 components of usable data set
#1
#2
#3
#4
#5
SD 4.0883 2.2309 1.66807 1.22893 1.10492
PV 0.5392 0.1605 0.08976 0.04872 0.03938
CP 0.5392 0.6997 0.78948 0.83820 0.87758
(NB) SD: Standard Deviation, PV: Proportion of Variance, CP: Cumulative Proportion
We have performed dimensionality reduction

applying PCA and obtained the below results
in order to classify factor loading over threshold limit within 5% significant level of test for
correlation coefficient. The resulting classifier
contains the following characteristic parameters:
2
So-called curse of dimensionality refers to various phenomena that arise when analyzing and organizing data in highdimensional spaces.
ISBN: 978-1-941968-38-32016 SDIWC
LtNo
LtAc
LtBr
LtOv
Driving Characteristics by k-NN

LtNo
LtAc
LtBr
LtOv
43(55.1%) 14(18.9%) 16(19.5%)
6(8.1%)
10(12.8%) 47(63.5%) 10(12.2%) 12(16.2%)
13(16.7%)
6(8.1%)
48(58.5%) 8(10.8%)
12(15.4%)
7(9.4%)
11(13.4%) 48(64.9%)
After comparison between Table 5 and Table

2, it is confirmed that recognition rate of Table 5 has been improved than one of Table 2
and PCA-based dimensionality reduction contributes to precision improvement of classification.
Secondarily, at SVM application to recognition
of driving behavior, we have obtained good
recognition rate(R = 60.13%) with gamma
value: = 0.2511886 and cost value: C =
3.981072 and Table 6, which is a result of
average of rate for right answers based on
the condition with the above good recognition
69
rate(R = 60.13%). Just like the case of kNN application, it is confirmed that recognition rate of Table 6 has been improved than one
of Table 3 and PCA-based dimensionality reduction contributes to precision improvement
of this classification, too.
Table 6. Improved Classification of Driving Behaiviour
LtNo
LtAc
LtBr
LtOv
4.4
Driving Characteristics by SVM

LtNo
LtAc
LtBr
LtOv
66(83.5%)
1(1.3%)
1(1.3%)
0(0.0%)
0(0.0%)
74(93.7%)
6(8.0%)
6(7.7%)
5(6.3%)
1(1.3%)
61(81.3%)
1(1.3%)
8(10.1%)
3(3.8%)
7(9.3%)
71(91.0%)
Consideration
Although it is not so sufficiently re-cognitive

rate for us to utilize for any case, we have
applied our result(classifier from improved
SVM recognition scheme) into recognition of
drivers behavior at the normal situation. Recognized result shows Table 7. According to
Table 7, it is confirmed that our classifier can
recognize that drivers characteristics, namely
driving tendency, is to make a left turn at the
traffic intersection with acceleration.This determination about recognition of driving behavior can be very reasonable and feasible because almost all drivers of middle age had been
previously taught to make a left turn with acceleration in their driving school days.
Table 7. Result Comparison of Drivers Right Turns
Patterns of Right Turns
PatternNo1[LtNo]
PatternNo2[LtAc]
PatternNo3[LtBr]
PatternNo4[LtOv]
Times (Percentage)
23 (25.0%)
39 (42.4%)
12 (13.0%)
18 (19.6%)
CONCLUSION
This paper describes our system configuration of data transmission between smart devices and Web server with SQL database facility, data mining application about driving
characteristics recognition based on machine
learning such k-NN and SVM, and practical
ISBN: 978-1-941968-38-32016 SDIWC
recognition/classification results according to

reduced/improved machine learning.
With application of PCA to dimensionality reduction, 36 dimensions of data set can be eliminated into 14 ones and rate of recognition
can be improved from 52.73% to 58.16% at
the case of k-NN as well as from 55.94% to
60.13% at SVM.
Our classifier based on improved SVM machine learning can recognize that the relevant
driver has characteristic tendency to make a left
turn with acceleration at the traffic intersection.
REFERENCES
[1] N.Lane, E. Miluzzo, H. Lu, D. Peebles, T.
Choudhury, A. Campbell, A Survey of Mobile
Phone Sensing, IEEE Communications Magazine,
Vol.48, No.9, September 2010, pp.140-150.
[2] A. Hofleitner, R. Herring, A. Bayen, Arterial travel
time forecast with streaming data: A hybrid approach of flow modeling and machine learning,
Transportation Research Part B: Methodological,
vol. 46, no. 9, November 2012, pp. 1097-1122.
[3] M. Fazeen, B. Gozick, R. Dantu, M. Bhukhiya,
M.C. Gonzalez, Safe Driving Using Mobile
Phones, IEEE Transactions on Intelligent Transportation Systems, vol. 13, no. 3, September 2012,
pp. 14621468.
[4] J.-H. Hong, B. Margines, A. K. Dey, A
smartphone-based sensing platform to model aggressive driving behaviors, Proceedings of the
SIGCHI Conference on Human Factors in Computing Systems(CHI 14), 2014, pp. 40474056.
[5] E. Shmueli, V. K. Singh, B. Lepri,A. Pentland,
Sensing, Understanding, and Shaping Social Behavior, IEEE Transactions on Computational Social Systems, vol.1, no.1, March 2014, pp. 22 34.
[6] M. Zhang, A.A.Sawchuk, Motion primitive-based
human activity recognition using a bag-of-features
approach, Proceedings of the 2nd ACM SIGHIT
International Health Informatics Symposium (IHI
12), 2012, pp. 631640.
[7] D.M. Karantonis, M.R. Narayanan, M. Mathie,
N.H. Lovell, B.G. Celler, Implementation of a
Real-Time Human Movement Classifier Using a
Triaxial Accelerometer for Ambulatory Monitoring, IEEE Transactions on Information Technology
in Biomedicine, Vol.10, No.1, February 2006, pp.
156 167.
70
Support and Improvement of the Educational Process in Regional Education in

Slovakia Through Tablet Classrooms
RNDr. Daniel Polin, PhD., Ing. Hedviga Petrukov, PhD., Jacek Stando,
Department of Informatics
Faculty of Education Catholic University in Ruomberok
Hrabovsk cesta 1,034 01 Ruomberok Slovakia
Lodz University of Technology, Poland
E-mail: polcin.daniel@gmail.com, hedviga.palasthy@ku.sk
ABSTRACT
Today, the use of information and communication
technologies and equipment is for teaching
necessity. This fact is mainly due to the time in
which we live and the progress that cannot be
stopped. Therefore, teachers must lead students to
use information and communication technologies
and their correct use in life. We can say, that the
innovations for schools are tablets.
The article analyzes the current state of
digitization of primary and secondary schools in
the Slovak Republic. It deals with the national
project: Digitization of the educational system of
regional schools, running in 2013-2015, and with
the introduction of electronic services in education
in the educational process. It acquaints readers
with the tasks and aims of digitization, it refers to
the issue of interrelated projects: Planet of
knowledge, Digtal school, School on touch.
At the end there is demonstration some suitable
applications for teaching physics using various
sensors supplied tablets with Android operating
system and other applications suitable for teaching
on tablets
KEYWORDS
national project, digitization, digital content,
tablets, tablet classroom
1 INTRODUCTION
The national project Electronising of the
educational system of regional schools should
carry and build a functional electronic
education system in Slovakia and introduce
electronic services to the operation. It also
includes the establishment and equipment of
digital classes, the customization of digital
content and finally the training of selected
staff to ensure the further education of the
pedagogical staff. It is financed from EU
structural funds under the operational
programme Informatisation of society and
ISBN: 978-1-941968-38-32016 SDIWC
took place in the period from Nov 22, 2013 to

Sep 30, 2015.
The schools across the country under the
project acquired modern digital equipment
which allows them to make better use of
digital educational content and modernize the
teaching in lessons. It covers over 5,680
interactive whiteboards, 5,680 laptops, 2,686
colour printers, 20,000 tablets and 1,000 WiFi routers in total. [3]
The project partner is the Methodological and
Pedagogical Centre, which is responsible for
organizational and staffing logistics of the
project and the use of digital educational
materials for modern forms of teaching
2
DIGITAL EDUCATIONAL
CONTENT
The digital learning content is an important
tool and an aid for the teachers to facilitate
their work and increases the attractiveness of
the learning for the pupils. Its implementation
is a logical step at present based on
widespread use of the information and the
communication technologies with the
interactive character. [1]
Making the digital educational content the
possibilities for teachers will significantly
improve in the field of visualization of more
complex processes, of actively involving
students or connecting theoretical knowledge
with practice. Thus it represents a very
convenient and useful supplement to existing
printed textbooks. Compared to them,
however, it can be much easier to correct or
update.
2.1 Planet of knowledge
Planet of knowledge is the first major project
of digitization of educational content in line
71
with the concept of the informatisation in

school sector with a view to 2020 DIGIPEDIA 2020.
The educational portal Planet of knowledge
represents a comprehensive tool for schools
and their teachers which is used for the
preparation the learning materials to work
with pupils during a lesson and then to check
homework assignments completed by
students.
The portal contains more than 30,000
educational materials in mathematics,
physics, chemistry, biology and science. The
educational materials contain multimedia
content in the form of videos, animations,
simulations, presentations, illustrations, 3D
models, images, photographs, interactive
exercises and lessons. The materials on the
portal are accessible by subjects and levels as
well as by thematic units defined by state
educational programme of the Slovak
Republic. The educational materials are
certified and recommended by the Ministry of
Education, Science, Research and Sport of the
Slovak Republic. [2]
2.2 The national project Electronisation of
the educational system of regional
schools - Digital school
The national project Electronisation of the
educational system of regional schools (EES
RS) is another step in fulfilling of the Concept
of the informatization of the education. The
subject of the project is the Ministry of
Education of the Slovak Republic, the project
partner is the Methodical and Pedagogical
Center,
which
is
responsible
for
organizational and staffing logistics of the
project and the use of digital educational
materials for modern forms of teaching. The
project is funded under the operational
programme Informatization of Society.
The current trend in the market of the
information and communication technologies
(ICT) is a global shift away from the use of
PCs and laptops to a more massive use of
tablet devices in all areas of the actual
situations.
In the area of development of the latest
technology to support teaching this trend
ISBN: 978-1-941968-38-32016 SDIWC
reflects in a shift towards the use of digital

educational content (DEC), i.e. processed
quality multimedia content provided by
modern devices and peripherals such as touch
tablets and interactive whiteboards which
involve a wider spectrum of sensory functions
of individuals. An attractive form of learning
by using tools and aids of DEC provides
opportunities for effective teaching (clearness,
connection of the reality with practical theory,
etc.).
The trend of growth in the use of modern
equipment has an impact on the use of
eGovernment services, where it is expected
that in a few years there will be a significant
part of the transactional eGov services
conducted
from
modern
mobile
devices. Therefore it is important to prepare
the next generation for the need of increased
digital literacy from school and pre-school
age.
These needs and trends have been in recent
years responded by the state which partly
invested for the creation of digital content
adapted to the existing infrastructure of
facilities in schools. However, it has not
supported the investments in modern terminal
equipment for modern teaching, whereas the
global trend is exactly the deployment of such
equipment. The investments were made
mainly in lower and upper secondary
education; there is a lack of coverage of the
pre-primary and primary education.
At present, kindergartens, primary and
secondary schools have insufficient IT
equipment. The vast majority of the
equipment used for ICT is morally obsolete
and very difficult to use for digital education
which is currently a priority in modern
education.
In the framework of the national project EES
RS there will be created an information
system called Electronic Services of the
Education System of Regional Schools
(ESESRS), which will provide an electronic
support of the processes aimed at creating and
making available of School Educational
Program (SEP). Such a system will form the
basis for the development of the curriculum
preparation for teaching, for the provision of a
72
digital curriculum as well as of the electronic

support for the actual teaching and home
preparation of pupils.
The educational process has been entered by
digital classes which represent a complete
solution to support the deployment of
elements of digital learning in all phases of
the educational process - preparing lessons,
actual teaching process and subsequent home
preparation. Provided educational content
continues and will continue to be tailored to
the modern end devices with touch control to
ensure a smooth transition from traditional
forms of learning to the teaching using digital
elements. [4]
For all kindergartens, primary and secondary
schools of a target area meeting the indicators
specified by the projects there are provided
sets of equipment containing an interactive
whiteboard with a laptop. For kindergarten
there are also provided educational packages
with educational software for school readiness
with digital educational materials and a color
printer. For primary schools there are
delivered educational packages to interactive
whiteboards for teaching mathematics on the
first level for a minimum of 15 selected
thematic units according to the state education
program ISCED1. [10] For selected schools
tablet sets with router equipment for a digital
classroom come. For teaching with tablets,
there are also adapted 250 educational units
for mathematics, physics, chemistry and
biology.
The information system of ESESRS will
serve the needs of educators, expert public,
students and their parents. All users will be
logged into the system by user names and
passwords and will have access to their
modified content both at school and within a
home preparation.
In terms of teachers the information system
will allow:
- searching for educational materials
and digital learning content suitable
for inclusion in the educational
process
- implementing
a curriculum
preparation and a preparation for
ISBN: 978-1-941968-38-32016 SDIWC
teaching directly using the elements of

digital learning
using system tools to manage the
lesson in a digital class subsequently
the home preparation using the
elements of digital learning [4].
In terms of students the system will bring the

traceability of digital materials used in
teaching and subsequently the use of these
additional materials within the home
preparation.
From the perspective of parents the
implemented information system will bring
the possibility of monitoring the progress of
the child in the learning process and within
the home preparation it will provide the
possibility of closer attachment to the topics
being taught at school.
In the area of school readiness the
implemented information system will provide
an access to a unified methodology and to the
tools for preparing and assessment of school
readiness of the staff for kindergartens and
primary schools, external experts and parents.
School staff and external experts thus will
receive a tool for a more effective preparing
and assessment of the development of
children.
The system will enable parents to monitor the
process of development of each child's ability
thereby it will also give the opportunity to
address areas of need within the child's home
preparation.
Initial experience suggests that students get
activated by touch technologies, which bring
significant motivational potential and create
favorable conditions for the learning process
2.3 Project School on touch
This is the first commonly known project
which had the task to introduce modern
didactical tools to schools - tablets. The
project was developed under the auspices of
the nonprofit organization EDULAB. The
project started in 2013 and worked until 2015.
Project in cooperation with Samsung has
equipped 15 schools with tablet classrooms
and with personal laptop for the teacher.
73
Fifteen selected schools received more than

four hundred tablets Samsung Galaxy Note
10.1 These are high-tablets with a special
touch screen to write whit the integrated pen.
The tablets will be through Samsung solutions
connected to school blackboard with touch
screen with a diagonal 65 "or 162.5 cm. Each
teacher got Samsung laptop for home use. All
technologies will remain after the project
schools
There were nine elementary schools, three
high school and three universities of the 15
schools. The project also has an educational
section "Technology is just the first step to
transforming education. Follow-up activities
of teachers and pupils have the technology to
revive them with his abilities. During the
project, pupils and teachers produce materials
that are published on the website
www.skolanadotyk.sk in the Materials." This
means that at the website are available
methodologies, models, teaching video blogs,
pupils' projects and creative workshop
EDULAB is a nonprofit organization that is
behind the school project to the touch. It
offers a variety of courses for teachers and
courses how to use the tablet at the hours.
Also the so-called center EDULAB was
created. It is the prototype of a modern
multimedia class as it could look like in the
future.
"The main goal of the EDULAB center is to
increase the popularity use of modern
educational technology in schools and show
the benefits and importance for the
modernization of education. EDULAB brings
teachers the opportunity to gain practical
skills by completing courses and try out with
your pupils work with the latest technologies.
" They also organize various competitions. [7]
3 "TABLET" TRAINING
As part of the training teachers learned to use
Samsung School solution to create working
materials or collective tasks. They trained
how to lead students' work with tablets during
school lessons. They prepared model lessons
with their students. [6]
The trainings called "Management of tablet
classroom" were implemented through the
ISBN: 978-1-941968-38-32016 SDIWC
cooperation
of
national
projects
Electronisation of the Educational System of
Regional Schools (Digital school) and
Activating Methods in Education.
3.1 School final projects
All schools involved in the project (both
tablet and non-tablet ones) committed in their
applications to the project to the development
of so-called "school projects". The schools
were committed to develop the following
projects: [5]
Mandatory topic for all involved primary and
secondary schools:
- 1 video, a multimedia presentation and
a methodical material to the topic: We
like Slovakia
Optional topic one of the following:
- 1 video and a methodical material to
the topic: Video experiments
- 1 video and a methodical material to
the topic: I learn, I teach you
- 1 dance video in the competition The
2014/2015 SCHOOL DANCE
3.2 Tablets Samsung Galaxy Note 10.1
Under the project Digital School, there were
delivered to schools tablets with the following
parameters:
CPU: ARM Cortex A9, quadcore 1.4

GHz processor
Display: 10.1" WXGA TFT,

resolution 2560 x 1600
Android 4.1.2
front and rear camera
flash and light sensor
microphone, speaker
IrLED
head phones connector and
volume button
Slot for memory card Micro SD
S Pen
Figure 1 Tablets Samsung in teaching
74
3.3 Google Play

The attractiveness of smart phones and tablets
lies in the option of installing a wide variety
of applications that the user finds in an online
store with both free and paid apps - Google
Play. Currently on Google Play a vast
plethora of application is available about 1.5
million, substantially greater part of them is
free.
4
Figure 3 Measurement of dimensions using photo and

reference subject to use rear camera with a flash option
SAMPLE OF SOME SUITABLE

APPLICATIONS FOR TEACHING
PHYSICS
These applications use a marvellous range of

quality sensors supplied tablets with Android.
They can replace with high quality many
older tools previously used in laboratory
sessions and then each student can realize
experimental measurements in parallel with
other students on tablet. [11], [12]
Figure 4 Distance meter allows you to calculate the

distance around the object by mathematical algorithms
to use the back tablet camera
Figure 5 Sound meter using

a sensitive microphone and
speaker
Figure 2 Bubble level. Uses a gyroscope OIS to

determine the gyro tilt and rotation tablet.
Figure 6 Goniometer measures the angle of inclination

e.g. inclined plane, the slope It uses a gyroscope and
motion sensor - an accelerometer
ISBN: 978-1-941968-38-32016 SDIWC
75
Figure 9 Speedometer measures by GPS immediate,

maximal and average speed of the vehicle, bicycle,
boat and so on.
Figure 7 Rangefinder measures the distance and height

of the destination using trigonometry and uses a rear
camera again with the possibility flash
Figure 10 Vibration meter uses sensors to measure

vibration or the earthquake
Figure 8 Light intensity meter measures light intensity

in lux and color temperature of light in Kelvin using
the light sensor
ISBN: 978-1-941968-38-32016 SDIWC
Figure 11: Use of tablets in the classroom Laboratory

of Physics
76
OTHER APPLICATIONS SUITABLE

FOR TEACHING AT THE TABLETS
The application is demanding on hardware, with

large maps there may be slight microlags.
5.3 Clarisketch
5.1 Socrative
Socrative is a product of Mastery Connect. This is
an application designed for testing pupils. It can
be freely downloaded at the Google Play. The
application supports only Android operating
system what is a disadvantage for tablets running
Windows. The application is free for both teachers
and pupils disadvantages are small ads like in all
free applications. The only supported language is
English
It is a program that is used to create a narrative

drawing on a tablet or phone. The application is
only for devices with Android OS and is
downloadable for free on Google Play.
When you first start there is also a short user
manual but the application is very simple and
particularly intuitive. Clarisketch does not require
Internet access.
Figure 14 The work of student in the program

Clarisketch
Figure 12 Socrative
The application requires during its use access to

the Internet to synchronize with the teachers'
tablet. [8]
5.2 Mindomo
It is a program designed just for creating mind
maps and their presentation. It is possible to put
into it different images and it disposes different
types of transitions. Mindomo supports all
platforms and can be downloaded at any store
such as Google Play.
The program can be downloaded for free and it is
possible to create freely in it only 3 maps.
Mindomo contains wide range of languages. The
work with it is simple. It also offers a number of
already pre-prepared images, the option to import
them or use video [9].
Conclusion
Every day the teachers are looking for the
ways to make their learning more efficient
and more fun through the use of mobile
devices. Currently it is a number of advanced
tools which can be used to improve the
teaching in the classroom.
To sum up the key messages which the
project Digital School sends and its impact on
the teaching process, they are as follows:
1. Creating of an electronic educational
system and acceleration of the process
of implementation of electronised
services and digital technologies at
schools.
2. Creating of digital educational content
and its distribution in accordance with
the state educational program of the
Slovak republic.
3. Creating of eGOV services system,
i.e. creating and accessing of school
educational program, accessing of
digital educational content and an
electronic testing of childrens school
readiness. [6]
Figure 13 Mindomo
ISBN: 978-1-941968-38-32016 SDIWC
77
Each school with a tablet classroom, which is

involved in the project Digital School, has
gained Samsung School - solution for the
management of tablet classroom. We believe
that the process of digitization of education in
Slovakia will continue and bring a significant
modernization of the preparation and the
actual implementation of modern teaching
supported by touch technologies now
commonly used by students and pupils of all
ages.
[12] PERSSON , J.R.-ERIKSSON, U. 2016.

Planetarium software in the classroom. Physics
Education, Volume 51 (2016), Number 2. ISSN:
1361-6552
Acknowledgements
The article was prepared within the project
TA04031376 Research/development training
methodology aerospace specialists L410UVPE20. This project is supported by
Technology Agency Czech Republic.
REFERENCES
[1] The central information portal of the Ministry of
Education Slovak republic. Online:
<https://www.iedu.sk/digipedia/elektronizacia_rs/
Stranky/default.aspx>
[2] Planet of knowledge - information portal of the
Ministry of Education. SR Online:
<https://www.iedu.sk/digipedia/planeta_vedomost
i/Stranky/default.aspx>
[3] Why digital learning content.
Online:
<https://www.iedu.sk/digipedia/preco_dvo/Strank
y/default.aspx>
[4] Digikola - NP Electronising education system's
Regional Education.
Online: <https://www.minedu.sk/digiskola-npelektronizacia-vzdelavaciehosystemu-regionalneho-skolstva/>
[5] Digikola coming to school.
Online: <https://www.minedu.sk/do-skolprichadza-digiskola/>
[6] The training for eGov Services Creating the
Schools education program.
Online: <http://skoly.digiskola.sk/skoly/>
[7] Project School to touch Online:
http://www.skolanadotyk.sk/o-projekte.html
[8] SOCRATIVEVIDEOS.
2014.
Socrative
Overview Online:
https://www.youtube.com/watch?v=gzNrXd5Xpu
4
[9] MS. KELLY GARTLAND. 2014. Mindomo
Demo. Online:
https://www.youtube.com/watch?v=Q6SVnms6D
-w
[10] GUNAGA, J.: Use of freeware software in
maths and natural sciences in primary education
In: Studia Scientifica Facultatis Paedagogicae :
Universitas Catholica Ruomberok. - ISSN 13362232, Ro. 12, . 3 (2013), s. 227-234.
[11] MAZZELLA, A - TESTA, I. 2016. An
investigation into the effectiveness of smartphone
experiments on students' conceptual
knowledge about acceleration. Physics Education,
Volume 51 (2016), Number 5.
ISBN: 978-1-941968-38-32016 SDIWC
78

Proceedings of The Third International Conference On Computer Science, Computer Engineering, and Education Technologies (CSCEET2016), Poland, September 2016

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Proceedings of The Third International Conference On Computer Science, Computer Engineering, and Education Technologies (CSCEET2016), Poland, September 2016

Încărcat de

Drepturi de autor:

Formate disponibile

The Third International Conference on Computer

Science, Computer Engineering, and Education

Published by: The Society of Digital Information

Virtual Labs and Educational Software as a Tool for more Effective

ISBN: 978-1-941968-38-32016 SDIWC

using virtual laboratories especially in

ISBN: 978-1-941968-38-32016 SDIWC

2 ACTIVITIES THAT SLOVAK

Preparing papers and essays

Cooperation with other students of

Organizing time and work

Cooperation with other students from

Tab 2. They use computer on most or every

Humanities (history, geography,

Natural science subjects (physics,

Other subjects (practical and

Art subjects (art / music / dance /

Language (Slovak or Hungarian)

saving data and creating graphs

Using educational software, designed

Creating computer programs, macros

Tab 3. Slovak pupils learned following tasks at

Typing messages using

Voice communication, on-line

Decide where to seek for information

Posting comments to on-line profiles

Decide which information is

Posting pictures, photos or videos to

Searching information for study or

Using WIKI or on-line encyclopedia

Sort the information gained from the

Respond the questions from other

Check the credibility of information

Writing posts on own blog

Asking questions on forums or web

Creating or editing web pages

Gain the access to the information

Present information using computer

Using computers and internet in out-of-school

Creating and editing documents

Creating simple presentations

Creating multimedia presentations

Using graphic and drawing software

Using spreadsheets for calculations,

ISBN: 978-1-941968-38-32016 SDIWC

Watching downloaded videos or

Using the Internet to obtain news about

Reading reviews, comments about

Searching information about places

+ the percentage of students is significantly bigger

During the teaching process, the ICT can

spreadsheet. Its big advantage is a user

Figure 1: Concave mirror

Figure 2: Derivative of the function

4 VIRTUAL LABORATORIES AND

Figure 3: Shutterstock/Macrovector ([16])

One of the unique capabilities of the

ISBN: 978-1-941968-38-32016 SDIWC

3. to standard of scientific research are

teaching and professional development for

At a time of rapid changes in science and

6 USAGE OF THE VIRTUAL

ISBN: 978-1-941968-38-32016 SDIWC