Improved collaborative filtering recommendation algorithm based on user attributes and K-means clustering algorithm

Abstract

Aiming at the problem of poor performance of collaborative filtering algorithm on data sets with large sparsity, this paper proposes an improved collaborative filtering recommendation algorithm which integrates user attributes and K-means clustering. When considering user similarity, the weight of user attributes is introduced to reduce the impact of data sparsity on similarity calculation. Meanwhile, the characteristics of user's age, gender and occupation are concerned. At the same time, combined with K-means clustering, the algorithm can further improve the accuracy of the recommendation model.