A dynamic composition of ontology modules approach: application to web query reformulation. (1st January 2013)

• Record Type: Journal Article
• Authors: Ben Mustapha, Nesrine; Baazaoui-Zghal, Hajer; Moreno, Antonio; Ben Ghezala, Henda
• Is Part Of: International journal of metadata, semantics and ontologies. Volume 8:Number 4(2013)
• Page Start: 309
• Page End: 321
• Publication Date: 2013-01-01
• Subjects: composition of ontology modules -- composition -- modular ontology -- semantic similarities; Metadata -- Periodicals; Semantic Web -- Periodicals; Ontologies (Information retrieval) -- Periodicals; Data structures (Computer science) -- Periodicals; Information theory -- Periodicals; 005.74
• Languages: English
• ISSNs: 1744-2621

Abstract

As the semantic web emerges, the problems related to the management of ontologies are gaining relevance. Ontology management is considered as one of the main issues in artificial intelligence and knowledge engineering. Ontology modularisation aims at structuring ontologies to facilitate its reuse and maintenance. Compositional methodologies intend to integrate a set of modules into a larger ontology. This paper presents a dynamic approach for ontology modules composition based on semantic similarity measures between concepts on the web, and aiming to improve the resulting modular ontology's structure and content. The proposed approach takes place in three main phases, namely: cooccurrence graph construction of ontology modules, cooccurrence graph clustering of modular ontologies and hierarchical modules clusters construction. To evaluate our approach, we compare the obtained results with those of other compositional approaches in the literature, and we also show how these modular ontologies can be used to improve web query reformulation.