Improved Query Expansion Using Relevance Feedback and Haar Wavelet Transform for Information Retrieval

Abstract

Abstract As the web is growing day by day increasingly, retrieval of a document with high precision is becoming a challenging task for the search engine. Another problem that retrieval models suffer, is the vagueness of user queries. Query expansion is one of the popular methods used to improve the performance of the retrieval system. This technique suggests some additional terms and appends them to the original query so that the performance of the retrieval system gets improved. The proposed method uses relevance feedback (RF) and based technique which assumes top 'k = 30' documents are relevant then from this user selects the 'n'(n varies according to system response) most relevant documents as per his information need. The proposed method also uses wavelet transform to select the most informative terms from top 'k' documents. The proposed method improves both precision and recall. The MAP (Mean Average Precision) value of the proposed method is 0.3792 which is an improvement of 14.83%, 14.87%, 30.21%, 14.32%, and 27.29% on MAP value is observed with respect to Bo1-based query expansion, Bo2-based query expansion, Chi-Square based query expansion, KL divergence-based query expansion and original query respectively.