| Author | Sort Borort |
| Call Number | AIT Thesis no.IM-09-10 |
| Subject(s) | Classification, Faceted
|
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Information Management, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Thesis ; no. IM-09-10 |
| Abstract | The amount of digital information on the web is rapidly and exponentially growing. Those web contents need to be efficiently organized and presented in the form which can be easily discovered to ensure that the hidden important information can be retrieved. Facet classification and folksonomy (a.k.a social tagging) are two among the most popular classification techniques which have been used in knowledge and information organization practice. These techniques enable users to search, browse, and annotate the information item with keywords or tags. However some limitations such as inefficient integrated architecture between the two methods and folksonomy problems remain the major issue for improvement. In this thesis we develop an alternative architecture and system that integrate faceted search with semantic tagging feature. We exploit the concepts and semantic relationships of tags from various knowledge base corpora and ontology such as DBpedia and WordNet as well as tags co-occurrence statistic for creating conceptual tags suggestion and recommendation system. We also use such concepts and relationships to automatically classify user-added tags into different predefined facets. Finally, we conducted an experimental study to measure the performance and accuracy of the tag classifier. With this approach, the suggestion of conceptual tags is used instead of regular tags. By using this, folksonomy problems such as inconsistent use of tags, tag ambiguity, polysemy, and plurality are solved to a great extend. Moreover, the conceptual tags are classified and organized into facets which enable user to search and browse information based on concept of tag in faceted navigation. The result from our experimental study shows that our approach can classify tags based on the concepts and semantic relationship of tag with the precision of 74% and the accuracy of 93%. This finding indicates that our approach is a promising approach for conceptual tags classification. |
| Year | 2009 |
| Corresponding Series Added Entry | Asian Institute of Technology. Thesis ; no. IM-09-10 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information Management (IM) |
| Chairperson(s) | Janecek Paul; |
| Examination Committee(s) | Haddawy, Peter;Dailey, Matthew N. ; |
| Scholarship Donor(s) | RTG-GMSARN; |
| Degree | Thesis (M.Sc.) - Asian Institute of Technology, 2009 |