Alternate Query Construction Agent for Improving Web Search Result Using WordNet

Abstract

Traditional information retrieval systems lack consistent semantic description of information i.e. they fail to meet users' need due to lack of applying semantic identification to extract the information from the available information. Use of semantic equivalent of the user query will improve the efficiency of the search. In this paper, we propose a framework for semantic based information retrieval. Here we find the concepts that user specify in their query by analyzing the semantic equivalencies. The result which is a set of alternate queries to the main search query is then compared with the existing keyword based system's result. Then, according to the alternate queries' search results, the main queries result gets rearranged by assigning new weights. We further personalize the search and then re-rank the results on user preference. The proposed semantic retrieval model is combined with keyword based model to achieve completeness of the knowledge base. The model which we propose is helping to project the most relevant result URLs to the higher ranks.

Optimized web search results through additional retrieval lists inferred using WordNet similarity measure

“…Now what we got is a set of alternate queries. The investigation on search result improvement using this set could project some of the documents to the top ranks as suggested in [16]. But the problem with all these queries which formed through this method is that not all of them are relevant.…”

“…The re-ranking problem is given importance in recent times to attain high precision at the very top ranks. In this process, the components like Ontology, thesaurus and dictionaries are used to better understand the query [14,16].…”

“…Further used, POS to develop a classifier to classify the specific sentences. 6 Lioma & Ounis, Similar work has been suggested to construct alternate queries through finding the semantic meaning of the search keywords [16]. Using the exact keywords often fail to retrieve the information sought by the user.…”