Analyze Collaborative Search History for Online Query Suggestion by Applying Re-Ranking Algorithm

Web search engine pursued popularity by providing real time search results to the user while seeking of information. To investigate search result query suggestion has become the most important feature of web search engine. List of suggested queries provides the previously searched and related search results for the current search state of user. Information for any particular topic is searched by the user on the web is still in the form of queries (i.e. keywords). Keyword based query similarity faces the problem of semantic ambiguity. Graph based similarity method solves a problem of semantic ambiguity. Suggested queries are set of recommended (i.e. related) queries collected according to the similarity relevance value by analyzing search history. Dynamic clustering algorithm resolves a problem of K-Means by providing flexibility to increase the number of clusters as data object points increases. Suggested queries may vary their position in a query suggestion group, according to the relevance value calculated by relevance algorithm. Graph based similarity check method collects set of suggested queries for the currently searching query by analyzing collaborative search history. The Comparative result calculated with precision, recall and F-Measure show re-ranking algorithm gives higher accuracy than online query suggestion method.