Abstract:

Traditional collaborative filtering algorithms suffer from poor similarity computation and sparse data,which lead to inaccurate recommendation information.In this paper,a new hybrid collaborative filtering algorithm is proposed to improve the quality of recommendation.The method of similarity computing is improved by using the weighted method,and the new framework is based on the memory of the two kinds of collaborative filtering algorithm.Experiment with the data set provided by Netflix shows that the algorithm has better effect than the traditional collaborative filtering algorithm.

Key words: collaborative filtering;similarity;sparse data