Sunteți pe pagina 1din 5

SQIT 3033 Knowledge Acquisition In Decision Making

Individual Assignment
Name : Tan Ling wei Matrik No : 211780 Lecture Name :En. Izwan Nizal Mohd Shaharanee Group : A

Due Date : 6th-March 2014

TRANSPORTATION AGENCIES PROJECTS Data mining (or knowledge discovery) is a new data analysis technique that process different perspectives and summarizing it into business applications. In the construction domain however, the use of data mining has been extremely limited. Data mining usually requires the availability of a large database of previous cases to be analyzed. Therefore applications in the construction industry must be geared to those situations where such databases are readily available. This paper describes a research effort to explore a potential use of data mining in the construction industry. Real data about asphalt paving projects was collected from various IDOT (Illinois Department of Transportation) sources and analyzed using data mining techniques. The results indicate that data mining can provide information beyond the use of general statistical analysis. Various rules and patterns were derived from the original database, which could be applied to support decision-making. The limitations of data mining are also noted including the need to verify and test the discovered patterns. The application of data mining to a database containing data on construction asphalt paving operations projects was explored. The main purpose was to explore any relationship between relevant variables that might reveal hidden knowledge about the paving projects. Ideally, the most interesting relationships to be identified are those between project cost and other variables, traffic control and traffic control cost and other variables, contractors and any cost variables, as well as any other general relationship between variables undetected. Following is the table collection : Table 1 Attributes and their state
Attribute General issues Contract no. District County Location Project characteristics Type of project Distance No. of lanes Planned working days Logical Numerical Numerical Numerical Numerical Numerical Logical Logical Type

Actual working days DBE Volume of asphalt concrete Surface mix Superpave Time of day Traffic control issues Traffic control Total traffic control cost Contractor issues Contractors no. Name of contractor Contractor's bid Percent change in bid No. of unsuccessful bidders

Numerical Numerical Numerical

Logical Numerical/Logical Logical

Numerical/Logical Numerical

Numerical Logical Numerical Numerical Numerical

General issues include 4 attributes such as Contract number, District, County and Location. General issues are to distinguishing between the instances. There may be certain facts buried in the data that can reveal connections between the general issues attributes and some other attributes. While project characteristics include 10 attributes. One of these attributes is the type of project which describes the type of project being constructed, which can be: Surfacing, Resurfacing, Patching, Widening, or a combination of the two. Typical aspects of operations represented by Length/Distance, Number of lanes, Planned Working days, Actual Working days, Volume of asphalt concrete, Mixture and Superpave attributes. And, The Volume is represented as QC/QA tons (quality control and quality assurance), which is used as a rough approximation of the total asphalt concrete for each project. The Mixture attribute is used to identify the asphalt concrete mixture used for every project. The Superpave attribute (Superior Performing Asphalt Pavements) indicates whether the project was a Superpave project or not. This is a factor that can affect the performance and productivity of the contractor. Next, there are only two attributes in traffic control which Traffic control (Boolean attribute) and Total cost of traffic control. Following is Contractors issues, it consists of the Contractor (its name and bidding number), Contractors bid (bid

price), Number of unsuccessful bidders (for each bid) and Percent change from contractors bid. Note a couple of key issues relating to the implementation of data mining in the construction industry are important. First, there is an obvious lack of standardization in the construction industry as it relates to collecting and storing project and company specific data. This industry fragmentation significantly hinders the uptake of data mining techniques in practice. Furthermore, the way in which the information is stored in the construction industry is generally not very organized and patchy. For example, various pieces of information had to be collected from bulletins, reports, as well as electronic databases and then re-structured into one database in order to facilitate data mining. In order to overcome this problem, the construction industry can learn from the stateof-the-art in data mining in other industry sectors. There is a need for a unified data model for construction data. This unified data model would be similar to the current building product model utilized in the Industry Foundation Classes (IFC) but would focus primarily on construction specific data. We envision this data model to be organized in three layers; one layer would capture project-specific data such as cost, estimate, schedule and productivity. The second would capture company-specific data, such as profitability, bids and bonding capacity. These two layers would be obviously interrelated so that information can be indexed from one layer to another. The third layer would capture industry-specific data such as employment rates, industry wide productivity rates and financial ratios. Another issue to note is the importance of having procedures in place to decide on which aspects of the construction data are suitable for mining, and to develop rational basis for data mining. This procedure must include a system for evaluating the results by potential end-users such as project managers and upper level executives. The results of discussions with IDOT personnel were then reviewed and some of the new rules discovered confirmed existing perceptions, such as the trend for new highway surfacing projects to be completed within scheduled time. Other pieces of information were suspected but the data mining procedure placed accurate probability values to them. Several new rules were completely new and provided new insight into the data, such as the relation between bid price and working days.

Refrences http://www.anderson.ucla.edu/faculty/jason.frand/teacher/technologies/palace/datamining.ht m Soibelman L. (2000), Construction knowledge generation and dissemination. BerkeleyStanford CE&M workshop: Defining a research agenda for AEC process/product development in 2000 and beyond. Witten, I. H., Frank, E. (2001). Data mining: Practical machine learning tools and techniques with Java implementations. Morgan Kaufman, California. Leu S., Chee N., Shiu-Lin C. (2000). Data mining for tunnel support: neural network approach. Journal of automation in construction, Volume 10, Number 4, pp. 429-441(13). TCC, Two Crows Corporation (1999). Introduction to data mining and knowledge discovery. Third edition. Two Crows Corporation. http://www.itcon.org/data/works/att/2007_8.content.06891.pdf Hand, D.J, Mannila, H., Smyth, P. (2001). Principles of data mining. MIT press, Massachusetts Miguel F. (2002). URL: http://www.softlookup.com/ Nii O. Attoh-Okine, (1997). Rough set application to data-mining principles in pavement management database, J. Comput. Civ. Eng., Am. Soc. Civ. Eng. 11 (4) 231-237. Soibelman L., Hyunjoo K. (2002). Data preparation process for construction knowledge generation through knowledge discovery in databases. Journal of Computing in Civil Engineering, ASCE, 16 (1), 39-47.
http://ascpro0.ascweb.org/archives/cd/2009/paper/CPRT192002009.pdf

Adrians P., Zantinge D. (1996). Data mining. Addison-Wesley Longman, England. Cabena P. (1997). Discovering data mining: From concept to implementation. Prentice Hall, NJ. Han J. (2001). Data mining: Concepts and techniques. Morgan Kaufmann Publishers, San Francisco.

S-ar putea să vă placă și