Cédric Pruski

Adaptive Ontology-Based Web Information Retrieval: The TARGET Framework

Finding relevant information on the Web is difficult for most users. Although Web search applications are improving, they must be more intelligent to adapt to the search domains targeted by queries, the evolution of these domains, and users characteristics. In this paper, the authors present the TARGET framework for Web Information Retrieval. The proposed approach relies on the use of ontologies of a particular nature, called adaptive ontologies, for representing both the search domain and a users profile. Unlike existing approaches on ontologies, the authors make adaptive ontologies adapt semi-automatically to the evolution of the modeled domain. The ontologies and their properties are exploited for domain specific Web search purposes. The authors propose graph-based data structures for enriching Web data in semantics, as well as define an automatic query expansion technique to adapt a query to users real needs. The enriched query is evaluated on the previously defined graph-based data structures representing a set of Web pages returned by a usual search engine in order to extract the most relevant information according to user needs. The overall TARGET framework is formalized using first-order logic and fully tool supported.

Experimental assessment of the TARGET adaptive ontology-based Web search framework

Finding relevant information on the Web can be a complex task for most of the users. Although Web search applications are improving, they still need to be more intelligent to adapt to the search domain targeted by users, the evolution of this domain and users characteristics. In this paper, we present an experimental assessment of the TARGET framework for improving the relevance of the documents when users are searching the Web by using adaptive ontologies. This is done first by introducing the TARGET approach. We will briefly present the used ontologies and their ability to adapt to domain evolution. Then, We detail the TARGET tool used in our experimentations. This includes its architecture, its ability to carry out the ontology adaptation process as well as the way it searches the Web and ranks the returned results. Finally, we discuss the results obtained using the tool through the presentation of our case study devoted to the retrieval of scientific articles.