An Approach for Personalized Web Information Retrieval using Modified PageRank Method

Abstract

Search engines - a tool to retrieve the desired information from web are based on the link structure of the web to arrange the Web Search Results (WSR) for a given query. Study of available literature highlighted that only a few studies consider the user's rating of a web page in ranking WSR. Also, most of the user queries are short and ambiguous and results obtained do not satisfy their information needs. So, it is desirable to personalize the WSR. The present study proposes a novel method for rating web pages by implicit methods and a modification to existing inverted index is presented. A modified PageRank algorithm using proposed Inverted Recommendation Index is also discussed. A novel strategy for personalized re-ranking the search results based on Long Term Interest (LTI) and Short Term Interest (STI) is also presented.