基于CLSVSM的电影评分预测及其推荐应用研究  

Movie Scoring Prediction Based on CLSVSM and Its Application in Recommendations

作　　者：牛奉高[1] 王恩慧 徐倩丽 NIU Fenggao;WANG En’hui;XU Qianli(School of Mathematical Sciences,Shanxi University,Taiyuan 030006,China)

机构地区：[1]山西大学数学科学学院,山西太原030006

出　　处：《山西大学学报（自然科学版）》2020年第2期261-266,共6页Journal of Shanxi University(Natural Science Edition)

基　　金：山西省高等学校优秀成果培育项目(2019KJ004);山西省高等学校创新人才支持计划(2016052006);全国统计科学研究项目(2017LY04)。

摘　　要：随着电影网站用户数量以及电影数量的上升,用户评分数据变得极其稀疏,导致推荐系统推荐质量下降。针对这一问题,文章在传统基于项目的推荐算法(IBCF)基础上提出基于共现潜在语义向量空间模型(CLSVSM)的项目评分预测算法。文章先通过CLSVSM得到电影共现矩阵以及电影共现相对强度矩阵,然后利用电影之间的共现潜在关系对评分矩阵进行补全,在此基础上预测用户对未观看的电影评分,进而生成推荐。实验结果表明:与传统的IBCF推荐算法相比,CMLVSM_IBCF算法的均方根误差(RMSE)和平均绝对误差(MAE)分别下降17.7%和17.6%。新提出的算法计算出的电影之间的相似度更准确,有效地减小了数据稀疏性对推荐结果的影响,显著提高了电影网站的推荐质量。With the number of movie website users and movies increase,the user rating data becomes extremely sparse,resulting in a decline of recommendation quality in the recommendation system.To solve this problem,the paper proposes a project score prediction algorithm based on the co-occurrence latent semantic vector space model(CLSVSM)based on the traditional project-based recommendation algorithm(IBCF).The paper first obtains the movie co-occurrence matrix and the film co-occurrence relative intensity matrix through CLSVSM,and then uses the co-occurrence potential relationship between movies to complete the scoring matrix,on the basis of which the user is predicted to watch the unwatched movie.Grading in turn generates recommendations.The experimental results show that the root-mean-square error(RMSE)and mean absolute error(MAE)of the CMLVSM_IBCF algorithm are decreased by 17.7%and 17.6%,respectively,compared with the traditional IBCF recommendation algorithm.The similarity between the films calculated by the newly proposed algorithm is more accurate,which effectively reduces the influence of data sparsity on the recommendation results,and significantly improves the recommendation quality of the movie website.

关 键 词：CLSVSM 评分预测 共现矩阵 相似性度量 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]