关键词: 个性化推荐, 协同过滤, 相似性

Abstract: Integrating classical collaborative filtering algorithm and network-based personal recommendation algorithm, this paper puts forward a new hybrid recommendation algorithm. It considers action relationship between objects in project based on the similarity and gets an improved hybrid algorithm: the MIX. Through the calculation of data set, the accuracy of the order is greatly improved, also the recommend accuracy.


Key words: personalized recommendation, collaborative filtering, similarity