摘要
协同过滤作为广泛应用的推荐方法,主要根据过去的交易和相似客户的反馈向当前客户推荐内容。但是,由于数据集稀疏性问题,很难区分客户之间的相似兴趣,这限制了协同过滤的可用性。针对数据集稀疏性问题,结合关联检索的相关知识,参考协同过滤算法,实现了一种改进的协同过滤算法来缓解稀疏性问题并提高推荐的质量。最后对提出的新算法进行测试,简述了改进方法的优点和不足。
As a widely used recommendation method, collaborative filtering mainly recommends content to current customers based on past transactions and feedback from similar customers.Due to the sparseness of the data set, it is difficult to distinguish similar interests between customers, which limits the availability of collaborative filtering.Aiming at the problem of sparseness of the data set, combined with the relevant knowledge of association retrieval and referring to the collaborative filtering algorithm, an improved collaborative filtering algorithm was implemented to alleviate the sparsity problem and improve the quality of recommendations.Finally, the proposed new algorithm was tested, and its advantages and disadvantages were briefly described.
作者
张洋
高艳华
郭晓坤
ZHANG yang;GAO Yan-hua;GUO Xiao-kun(The Second Academy of China Aerospace,Beijing 100039,China;Department of Software Research and Development,Beijing Institute of Control and Electronic Technology,Beijing 100038,China)
出处
《计算机仿真》
北大核心
2021年第9期495-500,共6页
Computer Simulation
关键词
协同过滤
关联检索
稀疏性问题
推荐质量
Collaborative filtering
Association retrieval
Sparsity problem
Recommendation quality
