Method of refining of document relevance in search engines based on a user's local behavioral description

Huge information content in the World Wide Web leads to redundancy of search result that can reach hundred thousands of pages. Due to this an end user sometimes finds necessary materials not in the top ten links. Method of local refining of document relevance is proposed in the paper. The proposed method is based on user's local behavioral description. Its main module comprises collecting the data from user's Web-surfing and monitoring the Web-resources retrieved in response to her/his web search query which are the most interesting for the user. Local search results are grounded on the base search engine results page and keyword search query. The research results indicate the 15% increase in search results quality.