Time Independent Query Recommendations Using Concept Based User Profile from Search Engine Query Logs
R. Umagandhi · 2014
2 Abstract: Search engines are highly confided resources of the people in mustering web information or obtaining relevant data approached for. The query log file contains an entry for every request posed by the user to the search engine and it is maintained in the system desktop or in the proxy server. Query log mining improves the performance of the search engine. The proposed algorithm mines the query log file which discovers the similar query keywords, URLs and the concepts based on both positive and negative preferences in its first phase. In the Second phase, the query cluster and the URL cluster is generated by using the combined similarity measure generated from the first phase. The cluster recommends the query to the user to frame their future queries based on their previous search histories and click through data. Expert's Query keyword and feedback are also considered for providing the recommendations. This approach also recommends the URLs to the user to be selected for their future queries.