An Algorithmic Query Refinement Model based on Query Classification

Objectives: This work focuses on developing an algorithmic refinement model for performing query refinement to improve the original query by adding more relevant candidate terms. The work has been tested in real time web environment and the results are also provided. Methods and Analysis: Query handling is a vibrant area in the field of Information Retrieval. Experiments reveal that formulating a query plays a vital role in generating relevant results. Since most of the users of web environment are naïve, query formulation cannot be expected to be effective always. This challenge could be overcome by the process of Query Refinement. Though there are many approaches put forward in the literature, in this work we propose a query refinement method based on the classification of queries and generating candidate terms from ontology and Thesaurus. Findings: We used the TREC 2014 web queries for evaluation and calculated Precision and Recall. The evaluation has been done on the Real time web environment which includes the most commonly used search engines like Google, Yahoo and Ask. The results have been compared and we could find a significant increase in both the precision and Recall. We did calculate the similarity measure using the Jaccard Similarity Coefficient. Since the Web environment is highly dynamic, the results tend to change depending upon the timing of execution of queries. Applications/Improvement: Addition of more relevant candidate terms helps to improve the accuracy of search applications. There has been no perfect retrieval system so far and there is always a scope of improvement. The result of this work reveals an increase in the accuracy of the retrieval system. We tend to analyze the results further, by comparing its impact in most of the available search applications.