融合阅读情绪的协同过滤文档推荐方法

Only Use attributes of the document as a recommendation basis cannot fully solve the problem since user's reading emotion about the similar document maybe affected after user has finished reading. This paper offered a new approach by combining user's reading emotion with collaborative filtering algorithm. We generated the factor that influence the reading emotion with user's reading time, and then, modified it according to the user's section in recommended list. Meanwhile, we use TF-IDF method to calculate the similarity between documents and re-ranking the recommended list using reading-emotion factor, Finally, the results of the tests show that the proposed algorithm fulfill the complicated needs of users.?