基于用户兴趣融合的协同过滤推荐算法

Collaborative filtering recommendation algorithm based on user interest fusion

协同过滤推荐算法已成功应用在各类门户网站,越来越多的研究者参与其中,然而在协同过滤推荐过程中用户—项目评分矩阵数据稀疏性以及推荐的准确性低等问题,始终制约着该算法的发展。为此,从用户兴趣角度出发,单独构建情景用户兴趣模型及社交网络用户兴趣模型,再通过Logit Boost算法将2个兴趣模型融合得到准确的用户兴趣模型。Slope One算法填充用户—项目评分矩阵,解决用户—项目评分矩阵数据稀疏性问题;同时,通过用户兴趣模型计算目标用户的最近邻居用户,通过计算用户相似性,得到推荐集来提高协同过滤推荐算法的推荐准确度。

Research on personalized recommendation based on user＇s interesting has been widely used in the field of recommender system, and become one of the more active research areas in the field of personalized recommendation, however, the sparsity of user project scoring matrix and low recommendation accuracy have always restricted the development of collaborative filtering recommendation algorithm. Therefore, from the perspective of user interest, this paper constructs the situational user interest model and the social network user interest model separately, and then uses the LogitBoost algorithm to fuse the two interest models for obtaining the accurate user interest model. Slope One algorithm filled user item rating matrix, solve the sparsity of user item rating matrix data. At the same time, the user＇s interest model is used to calculate the nearest neighbor user of the target user. By calculating the user similarity, the recommendation set is obtained to improve the recommendation accuracy of collaborative filtering recommendation algorithm.