改进的多因子协同过滤推荐算法

摘要

Compared with the traditional collaborative filtering recommendation algorithm, the tag based recommendation algorithm is more effective in mining users' preference model.But tag based recommendation algorithm cannot fully capture the dynamic changes of users' interest, for that it considers nothing about tags' timeliness.Therefore, this paper proposes an improved multi-factor collaborative filtering recommendation algorithm by combining the time factor of tags.The experiments are carried out on the data set MovieLens, the results show that the algorithm improved the accuracy and recall rate.%和传统的协同过滤相比,基于标签的推荐算法更能有效地挖掘用户的偏好模型,但传统的基于标签的推荐算法未考虑到标签的时效性,并不能充分捕捉用户兴趣的动态变化.因此,文中结合标签的时间因子,提出一种改进的多因子协同过滤推荐算法.在MovieLens数据集上进行实验,结果表明,该算法在准确率、召回率上均有所提升.