计算机应用研究2017，Vol.34Issue(5)：1397-1400,1414,5.DOI:10.3969/j.issn.1001-3695.2017.05.027
基于改进带偏置概率矩阵分解算法的研究
Research on improved bias probabilistic matrix factorization
摘要
Abstract
To solve the difficulty of data high-dimensional sparse problem in personalized recommendation,this paper proposed a new recommendation algorithm based on singular value decomposition and bias probability matrix decomposition.Firstly,the algorithm obtained the user-item rating matrix,then used the singular value decomposition to initialize the potential factor matrix of users and items.And further this paper integrated bias information with probability matrix decomposition algorithms to improve the accuracy of recommendation.Finally,it used maximum likelihood to transform the score prediction problem into an optimization problem by mini batch gradient descent.Simulation experiments on the three publicly available datasets show that the proposed algorithm can improve the recommendation accuracy on the three difference datasets,so as to ease the high-dimensional data sparse.关键词
概率矩阵分解/偏置/奇异值分解/个性化推荐Key words
probabilistic matrix factorization/bias/singular value decomposition/personalized recommendation分类
信息技术与安全科学引用本文复制引用
王建芳,张朋飞,刘永利..基于改进带偏置概率矩阵分解算法的研究[J].计算机应用研究,2017,34(5):1397-1400,1414,5.基金项目
国家自然科学基金资助项目(61202286) （61202286）
2015年度河南省高等学校重点科研项目 (15A520074) （15A520074）
