The Improvement of Association Rules to Use the Item-based Collaborative Filtering in the Recommender Systems

Although in today's world, the Internet put a great volume of data as an opportunity for the user to access, in the absence of efficient management on the large size of available data there will be hinder for improvement. Therefore, today given to the increasing volume of data and information, there is a need to have a system to lead the users to the considered product or service. Recommender systems present a method to create the personalized offers. One of the most important types of recommender systems is the collaborative filtering that deals with data mining in user information and offering them the appropriate item. Including the data mining methods is the generation of association rules that can be implemented through different algorithms such as FP-Growth. In this article using a data mining tool and FP-Growth algorithm application we have extracted the rules in two different dataset, and by investigating the support and confidence criteria we have assessed their efficiency. We also tried to improve the generated rules. The results can be used in collaborative filtering systems.

Keywords:

recommender systems , collaborative filtering , association rules , support , confidence