Multi-search: A Meta-search Engine Based On Multiple Ontologies

Abstract

In this paper, we present Multi-Search meta-search engine. Multi-Search combines three approaches: meta search, ontology-based semantic translation techniques, and statistically-based semantic relatedness measures. Multi-Search attempts to employ knowledge represented by multiple ontologies for both query translation and returned results merging. In addition, it utilizes semantic relatedness measures to address the issue of missing background knowledge in the used ontologies. The developed system operates on top of several search engines and can be easily extended. Experimental results indicate that the techniques used to build the meta-search engine are both effective and efficient.