Abstract

Paper Title/ Authors Name Download View

IMPROVED SEARCH RESULTS USING FEEDBACK SESSIONS

Kota Vamsee Krishna, Smita Deshmukh


Before submitting any query to a search engine, every user has a specific goal in mind but that goal is not known to the search engine. Based on the query related matching information which is currently available in the database, search engine will display the results and user will have to scan through them to find out the website of his interest everytime. From user experience perspective, this is a very time consuming and recurring activity. User would be happy if scanning several results to find out required information for a particular query can be reduced to a one time activity. User would practically benefit if search results could be displayed in categories (such as Business, Science and Technology, Sports etc.) with manual rating so that user defined highest rated URLs are displayed at the top which no search engine would do. It will help user to save effort and time in fetching required information. This can be achieved by implementing a desktop-based application. This paper is about a desktop-based application which will fetch search results from Google and Microsoft Bing together in one go and will infer user search goals. Users will have to register themselves and create user id and password. Considering one login/logout as a single session, usage logs will be captured and feedback sessions will be generated to restructure and optimize display of search results. The performance of this application will be evaluated using Classified Average Precision (CAP) factor.