Applying ontology based vector space model to web 2.0

Abstract

In social network the aggregation of RSS feeds leads to the syndication of users. Effective user oriented information retrieval can help decentralized users construct more effective social relations. In this paper the proposed ontology based vector space model (O-VSM) is applied to social network. It provides a novel approach for user-oriented information retrieval. RSS feeds are used as distributed information sources to illustrate how O-VSM works. RSS feeds are regarded as a set of knowledge fragments of decentralized users. In O-VSM, ontology can be expressed in graph and the semantic correlation of terms can be measured via the ontology graph. Under the support of the ontology, knowledge fragments can be transformed into local feature vectors that compose the ontology based vector space. In order to eliminate inherent semantic ambiguity and uncertainty, fuzzy inference is used to transform local feature vectors into global feature vectors in a uniform metric space. © 2008 IEEE.