摘要
个性化推荐系统中应用得最广泛的是协同过滤算法,而相似度的计算是协同过滤算法的核心。针对传统相似性度量方法中将用户对单个产品与单类产品的喜好未加以区分的不足,提出了一种基于用户兴趣与喜好的相似性计算方法。该方法根据用户兴趣与喜好,将对某个产品与某类产品的喜好程度区分开来,再通过加权的方式形成最终计算同类产品不同用户间的相似性。最后,采用Movie Lens数据集进行算法测试,测试实验结果表明,该计算方法的计算质量有明显提高。
The collaborative filtering algorithm is the most widely used in the personalized recommen- dation system, and the similarity calculation is the core of the collaborative filtering algorithm. In view of the deficiency of the traditional similarity measurement methods in the difference between the user preferences for a single product and a kind of product, a similarity calculation method based on user＇ s interest and hobby is proposed. According to the user＇ s interest and hobby, this method will be dis- tinguishing the user preferences for a single product and a kind of product. Then, the similarity be- tween different users of similar products is calculated by the weighted method. Finally, the algorithm is tested by the MovieLens data sets, and the test results show that the calculation quality of the proposed is obviously improved.
出处
《贵州师范大学学报（自然科学版）》
CAS
2017年第1期87-92,共6页
Journal of Guizhou Normal University：Natural Sciences
基金
国家自然科学基金资助项目(71161006)
贵州省高等学校基于知识的制造服务创新人才团队(黔教合人才团队字[2015]58)
关键词
个性化推荐
用户偏好
产品属性
相似度计算
personalized product recommendation
user preferences
product attributes
similarity cal- culation
