摘要 针对当前分布式潜在因子推荐算法存在时间复杂度较高、运行时间较长的问题,文中提出基于LU分解和交替最小二乘法(ALS)的分布式奇异值分解推荐算法,利用ALS利于分布式求解目标函数的特点,提出网格状分布式粒度分割策略,获取相互独立不相关的特征向量.在更新特征矩阵时,使用LU分解求逆矩阵,加快算法的运行速度.在KDD CUP 2012 Track1中的腾讯微博数据集上的实验表明,文中算法在确保一定推荐精度的前提下,大幅提升推荐速度和算法效率.
Abstract：Aiming at the problems of high time complexity and long running time of the current distributed potential factor recommendation algorithm, a distributed singular value decomposition recommendation algorithm based on LU decomposition and alternating least square(ALS) is proposed. Based on the characteristics of ALS for distributed solution of objective function, a grid-like distributed granularity segmentation strategy is proposed to obtain independent and unrelated feature vectors. When the characteristic matrix is updated, LU decomposition is adopted to solve the inverse matrix to speed up the operation of the algorithm. The experiment on Tencent Weibo dataset in KDD CUP 2012 Track1 indicates that the recommendation speed and efficiency of the proposed algorithm is significantly improved on the premise of ensuring a certain recommendation accuracy.
李琳, 王培培, 谷鹏, 解庆. 基于LU分解和交替最小二乘法的分布式奇异值分解推荐算法[J]. 模式识别与人工智能, 2020, 33(1): 32-40.
LI Lin , WANG Peipei , GU Peng , XIE Qing. Distributed Singular Value Decomposition Recommendation Algorithm Based on LU Decomposition and Alternating Least Square. , 2020, 33(1): 32-40.
[1] ZHANG S, YAO L N, SUN A X. Deep Learning Based Recommender System: A Survey and New Perspectives[C/OL]. [2019-07-25]. https://arxiv.org/pdf/1707.07435v3.pdf. [2] HE X N, HE Z K, SONG J K, et al. NAIS: Neural Attentive Item Similarity Model for Recommendation. IEEE Transactions on Know-ledge and Data Engineering, 2018, 30(12): 2354-2366. [3] 张志鹏,张 尧,任永功.基于覆盖约简的个性化协同过滤推荐方法.模式识别与人工智能, 2019, 32(7): 607-614. (ZHANG Z P, ZHANG Y, REN Y G. Personalized Collaborative Filtering Recommendation Approach Based on Covering Reduction. Pattern Recognition and Artificial Intelligence, 2019, 32(7): 607-614.) [4] 周 孟,朱福喜.基于邻居选取策略的人群定向算法.计算机研究与发展, 2017, 54(7): 1465-1476. (ZHOU M, ZHU F X. An Audience Targeting Algorithm Based on Neighbor Choosing Strategy. Journal of Computer Research and Development, 2017, 54(7): 1465-1476.) [5] 黄 璐,林川杰,何 军,等.融合主题模型和协同过滤的多样化移动应用推荐.软件学报, 2017, 28(3): 708-720. (HUANG L, LIN C J, HE J, et al. Diversified Mobile App Reco-mmendation Combining Topic Model and Collaborative Filtering. Journal of Software, 2017, 28(3): 708-720.) [6] DE PESSEMIER T, VANHECKE K, DOOMS S, et al. Content-Ba-sed Recommendation Algorithms on the Hadoop MapReduce Framework // Proc of the 7th International Conference on Web Information Systems and Technologies. Lisbon, Portugal: SciTe Press, 2011: 237-240. [7] LÄMMEL R. Google′s MapReduce Programming Model-Revisited. Science of Computer Programming, 2008, 70(1): 1-30. [8] ZHOU Y H, WILKINSON D, SCHREIBER R, et al. Large-Scale Parallel Collaborative Filtering for the Netflix Prize // Proc of the 4th International Conference on Algorithmic Aspects in Information and Management. Berlin, Germany: Springer, 2008: 337-348. [9] 余永红,高 阳,王 皓.基于Ranking的泊松矩阵分解兴趣点推荐算法.计算机研究与发展, 2016, 53(8): 1651-1663. (YU Y H, GAO Y, WANG H. A Ranking Based Poisson Matrix Factorization Model for Point-of-Interest Recommendation. Journal of Computer Research and Development, 2016, 53(8): 1651-1663.) [10] HUANG J, NIE F P, HUANG H, et al. Robust Manifold Nonne-gative Matrix Factorization. ACM Transactions on Knowledge Discovery from Data, 2014, 8(3). DOI: 10.1145/2601434. [11] 朱 夏,宋爱波,东 方,等.云计算环境下基于协同过滤的个性化推荐机制.计算机研究与发展, 2014, 51(10): 2255-2269. (ZHU X, SONG A B, DONG F, et al. A Collaborative Filtering Recommendation Mechanism for Cloud Computing. Journal of Computer Research and Development, 2014, 51(10): 2255-2269.) [12] 余志琴.基于ADMM的分布式矩阵分解.硕士学位论文.上海:上海交通大学, 2015. (YU Z Q. Distributed Factorization via ADMM. Master Dissertation. Shanghai, China: Shanghai Jiaotong University, 2015.) [13] 唐 云.基于Spark的大规模分布式矩阵运算算法研究与实现.硕士学位论文.南京:南京大学, 2016. (TANG Y. Research and Implementation on Large Scale Distributed Matrix Computation Algorithm with Spark. Master Dissertation. Nanjing, China: Nanjing University, 2016.) [14] CUI L Z, HUANG W Y, YAN Q, et al. A Novel Context-aware Recommendation Algorithm with Two-Level SVD in Social Networks. Future Generation Computer Systems, 2018, 86: 1459-1470. [15] PETRONI F, QUERZONI L. GASGD: Stochastic Gradient Descent for Distributed Asynchronous Matrix Completion via Graph Partitioning // Proc of the 8th ACM Conference on Recommender Systems, New York, USA: ACM, 2014: 241-248. [16] GIAMPOURAS P V, RONTOGIANNIS A A, KOUTROUMBAS K D. Alternating Iteratively Reweighted Least Squares Minimization for Low-Rank Matrix Factorization. IEEE Transactions on Signal Processing, 2019, 67(2): 490-503. [17] 谷 鹏.分布式群组推荐算法研究.硕士学位论文.武汉:武汉理工大学, 2017. (GU P. Research on Distributed Group Recommendation Algorithm. Master Dissertation. Wuhan, China: Wuhan University of Technology, 2016.) [18] GEMULLA R, NIJKAMP E, HAAS P J, et al. Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent // Proc of the 17th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York, USA: ACM, 2011: 69-77.
2020, Vol. 33 
