基于客户顾目的聚类协同过滤组合推荐算法研究

通过分析传统协同过滤存在的稀疏性、冷启动及实时性问题的根源后，提出一种改进的基于客户颅目的聚类协同过滤组合推荐算法；算法通过运用聚类技术和基于用户的协同过滤算法来预测计算邻居用户，并给出未评分的目标项目的最终预测评分以得到推荐列表，弥补协同过滤推荐在新项目推荐方面的不足的同时稀疏问题也迎刃而解；在预测评分中增加时间权重...

Researches of Collaborative Filtering Recommendation Algorithm Based on User and Item Clustering Combination

After analyzing cold-start, sparse and real-time issues in the process of collaborative filtering, we propose an improved collaborative filtering recommendation algorithm based on user and item clustering combination. By predicting the neighboring users with the use of collaborative filtering algorithm based on clustering technology and users, a recommendation list based on final rating scores of unrated target items is obtained, thus the shortcoming of collaborative filtering in the recommendation of new items is compensated, and simultaneously the sparse problem is solved. Then time weight is increased during predicting rating for the considerations of more weight for the latest user interest, thus the proposed algorithm can enhance the quality of recommendations.