融合项目属性和用户兴趣迁移的协同过滤算法

Collaborative Filtering Algorithm Combining Project Attributes and User Interest Transfer

协同过滤算法作为各种商业推荐技术最常使用的方法之一。然而,由于数据的稀疏性和用户的评分存在单一的相似性,低精度相似度度量降低了推荐系统的性能。针对上述问题,提出了一种协同过滤改进方法。方法基于项目的分类属性,以用户的兴趣度量用户的相似度,量化用户的兴趣所发生的动态迁移,构建新的相似度量模型。Movielens测试结果表明,提出的算法缓解了数据稀疏性,优化了最近邻的选取,与传统算法相比有着更高的推荐精度。

Collaborative filtering algorithm is one of the most commonly used methods in various commercial recommendation technologies.However,due to the sparsity of the data and the single similarity of the user’s score,the performance of the recommendation system is reduced by the low precision similarity measurement.Aiming at the above problems,an improved collaborative filtering method is proposed.Based on the classification attributes of items,the algorithm measures the similarity of users with their interests,quantifies the dynamic migration of users’ interests,and constructs a new similarity measurement model.Movielens test results show that the proposed algorithm alleviates the data sparsity and has higher recommendation accuracy than traditional algorithms.