International Journal of Social and Human Sciences 3 2009

Applying Ontology to Enrich Metadata for Learning Objects Discovery in Distributed Repositories
Shima Zahmatkesh, Ahmad A. Kardan, Amir Golshan

AbstractOne of the major challenges in e-learning development is search and discovery of an appropriate learning object among the distributed content repositories according to the needs and interests of the clients. SCORM presents a mechanism for share-ability and reusability of the learning objects, known as Sharable Content Objects (SCO's). Although SCORM presents some approaches for content reusability, but efficient searching process is a significant problem yet. Therefore, an effective searching mechanism is necessary for discovery and access to the required learning resources, for utilizing in the e-courses. But resource discovery within a heterogeneous collection of resources is a challenging problem. Different techniques like keyword-based search do not have required efficiency for searching the resources on the web. Semantic web has been proposed for resolving this problem. Different techniques Based on semantic data mining have been introduced. Some approaches like ontology were proposed to overcome heterogeneity. Ontology represents a set of concepts within a domain, and also the relationships between those concepts. W can enrich the information related to the learning objects, and develop a better searching methodology by using ontology for metadata of learning objects. In this work, according to our proposed ontology, an enriched metadata for a set of Learning Objects are shaped. By means of this kind of learning materials and their metadata our ontology could be evaluated for effective searching. KeywordsE-Learning, Learning Object, Metadata, Ontology, Sharable Content Objects (SCO). I. INTRODUCTION -LEARNING is a widespread methodology in education with more emphasis on the speed of learning process and its customization. For facilitating the construction of knowledge and skills in the learner, learning activities should be aimed [1]. One of the E-Learning goals to response this
F. S. Zahmatkesh is a researcher in Advanced E-Learning Technology Laboratory, Department of Computer Engineering and Information Technology, Amirkabir University of Technology, Tehran, Iran (e-mail: sh.zahmatkesh@aut.ac.ir). S. A. A. Kardan is the supervisor of Advanced E-Learning Technology Laboratory, Department of Computer Engineering and Information Technology, Amirkabir University of Technology, Tehran, Iran (phone:+982164542729; e-mail: aakardan@aut.ac.ir). T. A. Golshan is Postgraduate student collaborating with Advanced ELearning Technology Laboratory, Department of Computer Engineering and Information Technology, Amirkabir University of Technology, Tehran, Iran (e-mail: sh.zahmatkesh@aut.ac.ir).

necessity is wide access to learning resources with higher quality and lower cost. In this regard information technology has an important role to achieve E-Learning objectives. In order to improve access methods to educational information, different standards were created such as LOM and Dublin Core. The SCORM standard was introduced for reusability of learning contents and better management of the learning resources. The existing deficiencies of these standards lead to use semantic web and its technologies for effective learning. Semantic web mainly focuses on giving a welldefined meaning to resources, services, and information. For applying semantic web it is necessary to use ontology to describe resources and applications on the web. Therefore, rich metadata could be available by using ontology. By means of ontology, different projects have been developed in E-Learning domain. The CoAKTinG project [2] was developed to advance collaborative mediated spaces for distributed e-Science through the novel application of advanced knowledge technologies. The OntoEdue project [3] puts its emphasis on adaptability and personalization in learning by means of ontology. The EUME Onto [4] is an educational ontology that contains concepts of learning design, learning contents and learning resources. Weihong [5] proposed an Integrated Semantic E-Learning Platform which is an approach to integrate content provision, learning process, and learner personality. In another work [6] domain ontology was presented which is used for sharing content and services between repositories. The rest of this paper is organized as follows: first, in section II, we show the overview of Semantic Web, and metadata to establish necessary essential elements for the rest of the paper. In section III, we describe ontology and introduce some parts of our ontology in detail. Section V shows the process of creating metadata for learning objects, using the proposed ontology. In section VI we sketch out future works. And finally, in section VII the paper conclusion is provided. II. SEMANTIC WEB The semantic web is an extension of the World Wide Web in which content can be expressed semantically, and can be read and used by software agents. By getting semantic to contents, they could be found, shared and integrated more

750

International Journal of Social and Human Sciences 3 2009

easily. At its core, the semantic web comprises a philosophy, a set of design principles, and a variety of enabling technologies. Semantic web provides tools for knowledge representation and management, annotation of data and resources, discovery of services and resources based on their meaning and function, automatic composition of services, and inference over metadata and ontologies [1]. The main goal of proposing semantic web is to grant capability of analyzing all types of data including the content, links, and also transactions between people and computers. The semantic web architecture supports content with formal semantics. Thus, contents on the web can be discovered and used by automated agents. This will enable them to reason about Web content, and produce an intelligent response to unforeseen situations. Semantic web consist of different layers and use variety of tools and technologies like XML, RDF, RDF Schema, and OWL [10]. Each of these technologies is used in one of the layers of semantic web architecture. In view of the fact that XML lacks a semantic model and it has only a surface model, XML could not be the solution for propagating semantics through the Semantic Web. The Resource Description Framework (RDF) provides a suitable means for adding semantics to a document, but it seems to suffer from a lack of formal semantics for its modelling primitives, making interpretation of how to use them properly an error-prone process. A solution to this problem is provided by the third component of the Semantic Web, the ontology. We can use OWL for describing ontology. On the top of ontology layer there is a Query layer that gets information from databases. In upper layer, there is the Logic layer, which enables intelligent reasoning with meaningful data. Learning contents have some semantic annotation beside the main content. Thus for finding a desired content, we can use semantic querying and ontologies. As a result, semantic web can provide suitable platform for searching desired learning contents. III. METADATA Metadata is a data for describing another data, and is used to make an easy understanding process of data management and searching content resources. Metadata provides a common set of tags that can be applied to any content resource. Therefore, contents can be described, indexed, and searched, as a reusable content. In the E-Learning community different metadata standards are emerging to describe content resources like RDF, Dublin Core, and LOM [11]. Also different communities have developed their own metadata. Because of the variation and heterogeneity, different metadata can not work with each other. Creating a common understanding between terms in various metadata, vocabularies can be helpful. The most important issues for searching learning materials which are considered by the learners are [12]:

Content: What the learning materials about. Context: In which form learning material is presented. Structure: How a set of learning materials merge and create a learning course. Therefore, by using ontology in each of the above mentioned issues, both instructors and learners can get efficient results in respect to the designing and accessing courses, respectively. IV. ONTOLOGY Ontology is a data model that represents a set of concepts within a domain and the relationships between those concepts for representing and describing knowledge. Ontology suggests a formal description and common understanding of a specific domain [13]. It generally describes: Individuals. Individuals (instances) are the basic components of ontology. The individuals in ontology may include concrete objects, and abstract individuals such as numbers and words. Classes. Classes (Concepts) are abstract groups, sets, or collections of objects. They may contain individuals, other classes, or a combination of both. Attribute. It is properties, features, characteristics, or parameters that objects can have and share. Objects in the ontology can be described by assigning attributes to them. Each attribute has at least a name and a value, and is used to store information that is specific to the object it is attached to. Relations. Relations are the ways that objects can be related to each other. An important use of attributes is to describe the relationships between objects in the ontology. Typically a relation is an attribute whose value is another object in the ontology. Events. Events describe the changing of attributes or relations. For coding ontology, different languages have been created and the most common one is OWL. It is a widespread, expressive language that in terms of the ontology allows the use of external reasoning to compute the consistency of the model, classifying the ontology, query the model and retrieving individuals [14]. In this work we used Protg editor [15] to create and manage our ontology [16]. This editor provides a graphical view of classes, and a primary class called "Thing" which is the root class of all classes. The structure of learning objects was specified in the proposed ontology. In this structure, three elements were identified: Content Fragment, Content Object, and Learning Object. A Content Fragment is a content unit in its most basic form, such as text, image, audio, video, animation, table, chart, and so on. Navigational elements enable the sequencing of content fragments in a content object. Content Objects consist of some Learning Objects which cover a learning objective. These elements appear as classes in the proposed ontology.

751

International Journal of Social and Human Sciences 3 2009

SCORM presents a mechanism for share-ability and reusability of learning objects, known as Sharable Content Objects (SCO's). For implementation and evaluation of our ontology, Sharable Content Objects (SCO) was used as learning objects to attain reusability. In Figure 1 subclasses of Content Fragment could be seen. A Content Fragment could be text, image, audio, video, animation, table, chart, and so on.

Fig. 2 Content Objects

Fig. 1 Different Content Fragment

Educational Information consists of subclasses like Learning Style, Difficulty, Interaction, and Age Range. Technical Information is another class that consists of information like Format, Size, and Location. Information regarding Cost and Copyright is presented in Right Information. V. IMPLEMENTATION THE METADATA For evaluating semantic search by means of our ontology, a collection of e-learning contents is necessary. Different concepts of e-learning content domain have been introduced in the proposed ontology. Rich metadata for the e-learning contents was created by using these concepts. User as an instructor can use these concepts and a user-friendly interface to search variety of e-learning contents among distributed repositories on the web. To test our proposed ontology, the topic of "E-Business and E-Commerce" was selected to create an appropriate e-content. In this topic, different issues like E-Commerce Mechanisms, E-Commerce Transactions, Market Research and Online Advertising, and E-Commerce Support Services were introduced. The structure of learning objects which introduce in our ontology, consist of three elements: Content Fragment, Content Object, and Learning Object. According to our ontology, for the chosen topic different types of Content Objects like Description, Explanation, Example, Question, Self-assessment, Exercise, and Exam were created. For creating these Content Objects, different kinds of Content Fragments were utilized. Based on our ontology a Content Object is being made of some Content Fragments such as Text, Image, Video, Audio, Animation, Table, and Graph. In this work, a metadata of the set of Learning Objects was created. Protg editor was applied for creating metadata for

A Content Object represents the content of a Learning Object. In an e-course types of learning objects such as examples, questions and answers, exercises, descriptions, lectures, simulations, and others could be used as part of its content. As being illustrated in Figure 2, subclasses of a Content Object introduce different types of learning resources which are presented in learning e-content. As mentioned above the learning contents could be delivered to learners in different manners such as Description, Explanation, Example, Exam, Exercise, and Question and Answer. In the other part of the proposed ontology, metadata for learning resources were created. Different standards in learning resource metadata were reviewed. These standards are similar in some parts and different in other parts. On the basis of our studies on LOM and Dublin Core and after comparing them with similar ones, sufficiently a complete set of metadata were chosen. The most important subclasses related to metadata, including General Information, Educational Information, Technical Information, and Rights were designed. General Information includes information like Name, Identifier, Language, Keywords, Description, Date, and Creator. General information with the same title existed in most of the standards.

752

International Journal of Social and Human Sciences 3 2009

each of the Learning Objects. Beside the ontology created by protg, introducing instance for each of the classes of ontology is possible. Therefore, the metadata of the Learning Objects was formed in protg editor. The main class of our ontology is Learning Object. About 200 instances of this class were created for each of the files in our collection. As figure 3 showed Learning Object has different relationships with other classes. Learning Object has relationship with Content Object. This relationship could be used as metadata for instruction of Learning Objects discovery. Some of the relationships were related to the different information of the Learning Objects such as General Information, Educational Information, and Technical Information. Furthermore Learning Objects have different relationships with themselves which introduce the relation of Learning Objects with each others. According to these kinds of relationships, different kind of metadata could be formed for each Learning Objects.

Fig. 4 Metadata definition in protg

According to our ontology, each Content Object consists of some Content Fragments. Figure 5 showed the Content Fragments of "Description_27" which consists of three Content Fragments: Video_27, Text_27, and Image_27. As a result, "business_to_consumer_application" is a Learning Object containing Description and Example, and its Description consists of Video, Text, and Image.

Fig. 3 Learning Object relationships

Fig. 5 Content Object structure

A. Structural Metadata The first group in the metadata is related to the structure of the Learning Objects. This structure was described in our previous work [16] by details. Each Learning Object consists of some Content Objects which are created from combination of some Content Fragments. The different types of metadata which were created for a Learning Object are shown in figure 4. "business_to_consumer_application" is a Learning Object with two Content Objects: Example_28, and Description_28.

Therefore, user could specify the structure of the desired Learning Object in searching learning materials. For example, he could indicate that he needs a Learning Objects which describe a topic and give some examples for it. B. Information Metadata The second group of metadata includes different information about the Learning Objects. This information is classified in six categories. Table 1 shows this classification in details. The values of these metadata were defined for each of the Learning Objects. Figure 6 represent some of these metadata. User could search for desired Learning Object by defining the value for each of these metadata.

753

International Journal of Social and Human Sciences 3 2009

TABLE I INFORMATION METADATA CATEGORIES

class

General information

Educational information

Technical information

Right information

subclass Creator Date Description info Evaluation statement Identifier Keyword Language Name Scope Age range Difficulty Interaction Learning time Display duration Location Format Size Copy right Cost

Fig. 7 Learning Objects relationships

evaluated at AELT Group, in Amirkabir University of Technology, by putting these resources in different repositories around the campus, and conducting professors for searching desirable learning objects to create their own ecourse. Then, the consequence of this metadata in an efficient searching on the web will be calculated. VII. CONCLUSION The access to the desired content among a collection of different types of them is one of the important challenges for gaining e-content share-ability. Heterogeneity and lack of universal standard are the main problems in using distributed resources. To tackle these problems different solutions have been presented such as creating standards in e-content domain, and semantic web. Semantic web provides related semantics for data and content resources and helps us in searching content accurately, by creating sharable concepts in different domains. Ontology is one of the semantic web technologies being used in this study. The proposed ontology provides an enriched metadata for learning resources, and rendering a more accurate searching tool for discovery of learning content resources in the web empowered by the metadata. In this study, the structure of our ontology was utilized which was introduced in the previous work (Kardan, 2009). Creating a set of Learning Objects which have metadata is necessary to evaluate the proposed ontology. Therefore, in this work the process of creating metadata for Learning Objects was established. Thus by using Protg editor, a set of metadata was created. This metadata could be used as an OWL file or in databases. In the next step, using the metadata, a semantic search could be implemented and the efficient searching on the web could be calculated. ACKNOWLEDGMENT Hereby we would express our thanks to Iran Telecommunication Research Centre for the dedicated grant to this work under the contraction numbered T/500/20616, and dated on 18.March.2008.

Fig. 6 Metadata of Learning Object

C. Relational metadata The third groups of metadata associates to the relationships of one Learning Objects with others. In this regard, different kinds of relationships were introduced in our ontology. In this work, some of the relationships like prerequisite, post requisite, has part of, is part of, and proposed were utilized in creation of metadata for Learning Objects. It is mentionable that applying rules to these metadata, recommendation and introducing related Learning Objects is possible. Figure 7 shows an example of this kind of metadata. VI. FUTURE WORK In this paper, our ontology to improve metadata for learning resources was proposed. In the next step, using the proposed ontology a metadata for learning resources is created. In the next step the effectiveness of using this metadata will be

754

International Journal of Social and Human Sciences 3 2009

REFERENCES
[1] C. Allison, S. A. Cerri, P. Ritrovato, A. Gaeta, and M. Gaeta, "Services, Semantics, and Standards: Element of Learning Grid Infrastructure", Applied Artificial Intelligence, 2005, pp. 861879. K. R. Page, et al., "Collaboration in the Semantic Grid: a Basis for eLearning", Applied Artificial Intelligence, 2005, pp. 881-904. C. Guangzuo, C. Fei, C. Hu, and L. Shufang, "OntoEdu: A Case Study of Ontology-based Education Grid System for E-Learning", Journal of Global Chinese Society FOR Computers in Education, 2004. R. Amorim, M. Lama, E. Snchez, and X. A. Vila, "An Educational Ontology based on Metadata Standards", 2004. H.Weihong, W.David, W.Dawn, and I.Tanko, "An Intelligent Semantic E-Learning Framework Using Context-Aware Semantic Web Technologies", British Journal of Educational Technology, May 2006, pp. 351-373. Z.Xin-juan, L. Xian-Feng, and G. Wei, "Ontology Based Sharing and Services in E-Learning Repository", International Conference on Network and Parallel Computing ,2007. C. Ostyn, "In the Eye of the SCORM, An introduction to SCORM 2004 for Content Developers", 2007. C. T. Yang and H. C. Ho, "An e-Learning Platform Based on Grid Architecture", Journal of information science and engineering, 2005, pp. 911-928. G. Mackenzie, and M. Baeini, " A (Mostly) Painless Introduction to SCORM", 2004. http://en.wikipedia.org/wiki/Semantic_web W. Hodgins, and E. , "Draft Standard for Learning Object Metadata", Technical Report: Learning Technology Standards Committee of the IEEE, 2002. L. Stojanovic, S. Staab, and R Studer, "E-Learning Based on the Semantic Web", In World Conference on the WWW and Internet, Orlando, Florida, USA, October 2001. http://en.wikipedia.org/wiki/Ontology_computer_science G. Vega-Gorgojo, M. L. Bote-Lorenzo, E. Gomez-Sanchez, Y. A. Dimitriadis, and J. I. Asensio-Perez, "A Semantic Approach to Discovering Learning Services in Grid-based Collaborative Systems", Future Generation Computer Systems, May 2006, pp. 709-719. http://protege.stanford.edu/ A. A. Kardan, S. Zahmatkesh, "A Proposed Ontology for Effective Searching of Sharable Content Objects Emphasizing on Learning Objectives", In 6th International Conference on Information Technology : New Generations, Las Vegas, Nevada, USA. submitted for publication

[2] [3]

[4] [5]

[6]

[7] [8]

[9] [10] [11]

[12]

[13] [14]

[15] [16]

755