User Centric Web Page Recommender System Based on User Profile and Geo-Location

A web recommender system assists a user to manage the vast amount of search results suggested by a search engine when surfing the Web with the help of recommender system agents. In this case the recommender system will reduce the effort a user has to spend in searching and navigating to important web pages or web documents and simultaneously increases the user ability to find more relevant documents. When same query is used by different users in finding useful information on the World Wide Web (WWW), the search engines return the same result for all users regardless of who submitted the query. But in a real time situation each user may have different information needs for the same query. Therefore, the search results should be based on the user information needs. In this paper we propose a recommender model which recommends web pages to a user based on the user profile and geographic location of the user. We propose a web-recommender model based on user location and a personalized search method to improve the search strategy and the quality of the results. We first create a user profile, identify the user location and then return the result to the user based on user query considering the user profile and location of each user. After that the web is crawled moving from link to link and an index page is constructed that includes the user keywords used as a query in the search. This index references with a user profile and a user search query. The recommender system, then lists the pages that have the same keyword or related with the keyword that were written as the user's search terms and returns the ranked results in reference to the user profile and user location. Trying out multiple sites while looking for information is a basic concern in web page recommendation systems. The engine looks for keywords throughout each Web page, and returns the result based on the matching with the user profile, frequency of the key word on the page and based on the appearance of key words in the title of the web page or in the URL directly.