基于用户兴趣变化的隐语义协同过滤算法

PLSA Collaborative Filtering Algorithm Based on User Interest Change

协同过滤是推荐系统中广泛使用的算法.协同过滤模型没有考虑用户兴趣的动态变化,影响推荐质量.为提高推荐准确度,提出新的推荐算法——将基于动态时间窗口的协同过滤推荐与高斯概率隐语义模型结合,利用动态时间窗口捕捉用户的兴趣变化,并结合概率隐语义模型分析用户长期兴趣,进而为用户提供更准确的推荐.实验表明,该算法同其他协同过滤算法相比具有更高的准确度.

Collaborative Fihering（CF） is widely used algorithm in recommender system. Collaborative Filtering model does not consider the dynamic change of user＇s interests, influence recommendation quality. In order to improve recommendation precision, a new recommender algorithm was pro posed, which combines the Collaborative Filtering based on dynamic time windows that capture change of userrs interest by dynamic time windows, and the Gaussian probabilistic latent semantic analysis（PLSA） that capture user＇s long-term interest together. The experimental results on Movielens dataset show that the new algorithm compares favorably with other collaborative filtering algorithm.