A new collaborative filtering algorithm using K-means clustering and neighbors' voting

Abstract

The Collaborative Filtering is the most successful algorithm in the recommender systems' field. A recommender system is an intelligent system can help users to come across interesting items. It uses data mining and information filtering techniques. The collaborative filtering creates suggestions for users based on their neighbors' preferences. But it suffers from its poor accuracy and scalability. This paper considers the users are m (m is the number of users) points in n dimensional space (n is the number of items) and represents an approach based on user clustering to produce a recommendation for active user by a new method. It uses k-means clustering algorithm to categorize users based on their interests. Then it uses a new method called voting algorithm to develop a recommendation. We evaluate the traditional collaborative filtering and the new one to compare them. Our results show the proposed algorithm is more accurate than the traditional one, besides it is less time consuming than it.