Analysis of Desktop Search and Ranking of Their Results Based on Semantics from User Feedback

Abstract

This paper explores the ranking of search results in a desktop search engine. A comparison between web search and desktop search and the respective user behaviour during them, is made. This is followed by an analysis of the existing desktop search tools. The paper then discusses a Bayesian technique to collect user feedback and use it to rank the results of existing search frameworks. The results of evaluation of 449 query-result pairs are reported. It is shown that incorporating user feedback improves the ranking of results by as much as 7.32%.