Synergies Between Association Rules and Collaborative Filtering in Recommender System: An Application to Auto Industry

Recommender system, famous for finding potential requirement of customers, is widely applied in many domain, such as bank, mobile, music, book and so on. We propose an integrated recommender system contains data-processing, recommendation and evaluation processed. Traditional recommendation process aimed to recognize items that are more likely to be preferred by a specific user, however, it is expensive to recommend items to users who have no buying intention. Therefore, we propose a two stage recommendation process by adopting advantages of many recommender technology. The first stage use association rules as a means of classifying customers and finding potential customers. In the second stage, CF methods are applied to realize recommendation. In experimental study, we use auto industry database to illustrate the proposed system. First, we find some implied information for the rules generated, which conforms to the observation. Second, CF method based on users’ implicit preference information to recommend.