结合时间上下文挖掘学习兴趣的协同过滤推荐算法

提出了一种基于时间上下文的协同过滤推荐(TCCF-LI)算法,实现了基于高校图书馆图书借阅记录数据上的学生学习兴趣挖掘.在传统协同过滤算法上引入时间上下文信息,既考虑了大尺度用户群体爱好的趋同性,又兼顾了小尺度个体用户爱好的短时相关性,获得了更高的推荐性能.在实际数据集上的实验结果表明,该算法在推荐精准度、召回率等方面比传统推荐算法有较好表现.

A Collaborative Filtering Recommendation Algorithm with Time Context for Learning Interest Mining

A time context based collaborative filtering recommendation (TCCF-LI) algorithm this paper was proposed, and students' learning interest mining from university library borrow record was implemented. The time context information was imported into the traditional collaborative filtering recommendation algorithm, in which, both interest homoplasy of large scale user groups and short-term correlation of small scale user groups was considered. Good recommendation performance was gained. According to the experiments on real dataset, TCCF-LI algorithm presents higher precision and recall rate compared with traditional recommendation algorithm.