A novel algorithm for restructuring web search results by clustering pseudo documents

The search results produced by the search engine are in scattered form. The user may enter the same query for different information needs. So, it is difficult for user to get accurate information related to the submitted query. In this system feedback session is used to show the click sequences of user for a particular keyword. These click sequences are mapped into TF-IDF vector which is used to generate pseudo document. A Pseudo document represents the most important words for the searched keywords. User search goals are generated by using K-means clustering algorithm from pseudo document and the restructuring of web search results is done on the basis of user search goal. But, sometimes the URLs are not considered in clustering because of less TF-IDF values. This results in omission of some URLs from restructured result. So we propose new algorithm to restructure web search results. By using this algorithm the restructuring is done properly and all the URLS are present in the restructuring results. Finally, the metric VAP, risk and CAP are calculating the performance of restructuring. After conducting experiments on the system, we observed that the results were improved.