用户兴趣变化和类别关联度的混合推荐算法

协同过滤算法是目前推荐系统中最普遍的个性化推荐技术。针对传统算法相似性度量方法不足的问题，提出了融合用户兴趣变化和类别关联度的混合推荐算法。算法根据用户的评分项目信息来对项目进行类别划分，挖掘出用户对不同类别项目的喜爱关注程度；同时将基于时间的兴趣度权重函数引入项目相似度计算之中来进一步提高计算的精确度，最后将改进后的相似度计算方法融入到用户聚类方法中，用户聚类之后，其所在的类别将对用户推荐准确度产生极大的作用。实验结果表明，在Movielens-1k数据集上运行该算法，该算法在运行效率和精确度上都有所提高。

Hybrid recommendation algorithm for user interest change and category related degrees

Recommendation system has been widely applied to various types of e-commerce sites, which effectively solved the problem of information overload, collaborative filtering algorithm is the most common in the recommendation system of personalized recommendation technology. Based on the problem of the traditional method of similarity measurement, a hybrid recommendation algorithm is proposed to combine the change of interest and class correlation degree. The algorithm classifies the project according to the user''s rating project information, and finds out how much the user likes to pay attention to different categories of projects. At the same time, the time based interest weight function is introduced into the project similarity calculation to further improve the accuracy of calculation. Finally, the improved similarity calculation method is integrated into the user clustering method. After the user clustering, the category of its location will have a great effect on the user''s recommended accuracy. The experimental results show that the algorithm is improved in operation efficiency and accuracy in the moviels-1k data set.