Multi Agent based Page Ranking Algorithm for Personalized Web Search with Query Optimization

Abstract

Personalized web search becomes a hot topic in search engine optimization where the search engine has to provide user centric results for any query. There are many approaches have been discussed earlier and suffers with irrelevant results. We propose a multi agent framework, where the search query has been executed over many search engines to get more valuable results. The proposed approach maintains user logs about the query submitted, page visited, time spent, activities performed on the web pages. The agents of the framework transforms the query into more meaningful and submits to more than one search engine. The results fetched form multiple search engines is evaluated for their relevancy according to query and from the relevant results a set of more relevant results has been extracted. Each resultant web page has been identified with the topic of information it has and the user actions are monitored and recorded. The extracted results are re-ranked to form the final result according to many constraints like, visit history, time spent, actions performed and etc.