Survey on the Use of Feedback Sessions for Inferring User Search Goals

Abstract

Largest source of web traffic are search engines. Search engines are being used by different kind of users for different purpose. When users are searching something they have a different search goal in mind. Thus the queries are mostly ambiguous one. In order to improve search engine relevance and thus user experience inference and analysis of user search is required. To get the best results it is needful to capture different user search goals. This paper first talks about the different ways of inferring user search goals. Then insights of new approach has been discussed. A new algorithm firstly specifies a framework to analyze user search goals by clustering feedback sessions. There should be a proper way to represent these feedback sessions. In the second step of this algorithm pseudo-documents are prepared to represent feedback sessions. With this original results are restructured. This in turn is used to select optimal user search goals.