一种基于改进协同过滤推荐算法的电影推荐方法  

A Movie Recommendation Method Based on Improved Collaborative Filtering Recommendation Algorithm CHEN Ying,SHEN Yang

作　　者：陈瑛 沈阳 CHEN Ying;SHEN Yang(College of Information Engineering,Guangzhou Institute of Technology,Guangzhou 510075,China)

机构地区：[1]广州工程技术职业学院信息工程学院,广东广州510075

出　　处：《现代信息科技》2025年第5期105-108,114,共5页Modern Information Technology

基　　金：广东省高等职业教育教学质量与教学改革工程项目(2023JG048);广东省教育科学规划课题(2022GXJK552);广东省2022年度高等学校科研平台和项目重点领域专项(2022ZDZX1067)。

摘　　要：针对协同过滤推荐算法中相似度计算存在的误差问题,文章提出了一种带修正因子的相似度计算方法,并将其应用于电影推荐算法中。首先,对电影信息数据进行降维处理,并通过K-means算法进行聚类,以提高算法效率。其次,引入修正因子,结合皮尔逊相关系数计算电影相似度矩阵。最后,根据相似度矩阵计算用户对未打分电影的预测得分,从而完成推荐。实验结果表明,与传统协同过滤推荐算法相比,本算法的效率提升了62.4%;优化后的相似度算法比传统相似度算法的平均绝对误差平均降低了9.67%。Aiming at the error problem of similarity calculation in Collaborative Filtering recommendation algorithm,this paper proposes a similarity calculation method with correction factor and applies it to movie recommendation algorithm.Firstly,the movie information data is reduced in dimension and clustered by K-means algorithm to improve the efficiency of the algorithm.Secondly,the correction factor is introduced,and the movie similarity matrix is calculated by combining the Pearson correlation coefficient.Finally,according to the similarity matrix,the user's prediction score for the unrated movie is calculated to complete the recommendation.The experimental results show that compared with the traditional Collaborative Filtering recommendation algorithm,the efficiency of this algorithm is improved by 62.4%.The Mean Absolute Error of the optimized similarity algorithm is 9.67%lower than that of the traditional similarity algorithm.

关 键 词：相似度矩阵 修正因子 主成分分析 聚类 协同过滤 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]