JACK Intelligent Agent-Based and User Preference Mining for Collaborative Filtering

With the rapid development of the Internet and the worldwide popularity of the Web is growing exponentially. Automation collaborative filtering (CF) is becoming efficient tool to assist users with accurate information. We present multi-agent model based on collaborative filtering system to find similar users' interesting and choose JACK Intelligent Agentstrade to design CF system. Meanwhile, a novel mining frequent pattern algorithm, which is used to find users' interesting items and deduce association rules, is presented. Compared with traditional CF system based on frequent pattern mining algorithm, the novel algorithm has higher recall and accuracy rate.