A New Approach on Restructuring and Reranking of User Search Results by using Feedback Sessions

Priti Alok Mishra, Bidisha Roy

Abstract: To enhance the significance of web search engine, it is very essential to meet the users requirement of results. These user search results can be generated with the help of feedback sessions. The framework in this approach firstly creates Feedback session out of user click-through logs. The click-through data will later be preprocessed for precise information retrieval. The original search results will then be restructured and they would be re ranked according to users domain of interest. Lastly, to evaluate the performance of search results, Average Precision will be used. The main advantage of this framework is that it will provide the most optimum search result in short period of time. Also it will reduce those noisy URLs that degrade the performance of web search engines on large scale.

Keywords: Click-through logs, Feedback Session, Re ranked search results, Average Precision