Mária Bieliková > Teaching

Sharing Information in a Portal Based on the Semantic Web

Ladislav Rado

Master thesis project supervised by prof. Mária Bieliková

Motivation

Current data representation on the web is a problem for machine processing, which the semantic web tries to solve. The main aim of this thesis is to discover communities of researchers with common interests from the metadata in the domain of research publications. We use Web Ontology Language (OWL) for metadata representation and for describing semantic relationships. We have designed a method for discovering communities based on a combination of two methods: PageRank for ranking authoritativeness of publications and authors, and Probabilistic Latent Semantic Analysis for discovering communities using citation graph and classification of the publications. Communities of authors are presented in a portal which is based on semantic web in a textual and graphical form.

Keywords: semantic web, metadata, communities of researchers, citation graph

Results

Obtaining Metadata - we analyzed various datasets and created crawler which gathers datasets from ACM portal.
Preprocessing - we created a rule based citation parser which reconciles references.
Analysis of metadata, discovering communities - we implemented 2 methods for discovering communities based on probabilistic models for clustering documents and ranking of author's authoritativeness.
Presentation in the portal - we implemented a portal which uses semantic repository for storing all metadata about discovered communities. Portal presents communities in textual and graphical form.

Evaluation: Semantic Web portal

Discovered communities by proposed method in textual form:

Discovered communities by proposed method in graphical form:

Publications

Rado, L.: Sharing Information in a Portal Based on the Semantic Web. Master thesis, Slovak University of Technology in Bratislava 2009. pdf (in Slovak)