A Proposed Framework for Building a Recommender Search Engine

摘要

There are billions of web pages had been hosted in the Internet and the number of documents is increased overtime. But since internet is a huge resource, the problem of how to determine what users need to give good search results is a challenge for many search engines. This paper suggest a framework for building general purposed search engines which can recommend search results (web pages) to user by using a hybrid recommender engine from content-based and collaborative filtering. This framework determines the roles of topics, websites and web pages and give recommendations rely on the mutual relation of the triple topics, websites, and web pages. With this approach, we can improve search quality of search engines and solve some problems that restricted recommender systems. Experimental results show that our search results better than search results of a search engine apply classical Google page rank algorithm [13], a well-known page rank algorithm used in many search engines.