An Intelligent Information Retrieval System Using Automatic Word Sense Disambiguation

This paper aims to establish that an intelligent contextual information retrieval (IR) system can improve the quality of search results by retrieving more relevant results than those obtained with traditional search engines. Search engines capable of implicit, explicit, and no contextual retrieval were designed and implemented and their performances studied. Experimental results showed that search engines with contextual IR produce results that are more relevant, and the outcomes further indicate that there is no perceived gain in choosing specifically any one of the two approaches of implicit or explicit. The performance of the indexing mechanism, as it classifies document tokens with their appropriate contexts/word sense, was evaluated. The effectiveness of the word sense disambiguation process was found to depend to a great extent on the process (implementation) as well as the raw data (thesaurus).