Journals Proceedings

Abstract

Search engine transaction logs have been investigated through large number of studies. Here mining of search engine query log is done. By mining user’s feedback, ranks of search engine result pages are optimized, so that related pages come earlier in the list. In this way user can easily and quickly find desired pages. Here web mining techniques are used to order the documents. This method first mines query logs using a novel similarity function to perform query clustering of similar queries. Then it discovers sequential pattern of clicked URLs in each cluster using existing Sequential Pattern mining algorithm PLWAP. In the end, search result list is re-ranked by updating existing rank values of pages using discovered sequential patterns. By this method, user desired relevant pages move upwards in result list and reduce search space for users.