南京大学学报(自然科学版) ›› 2019, Vol. 55 ›› Issue (6): 952–959.doi: 10.13232/j.cnki.jnju.2019.06.008

• • 上一篇    下一篇

基于张量分解和深度学习的混合推荐算法

张家精1,夏巽鹏2(),陈金兰3,倪友聪4   

  1. 1. 安徽建筑大学数理学院,合肥,230601
    2. 安徽建筑大学电子与信息学院,合肥,230601
    3. 安徽建筑大学机械与电气工程学院,合肥,230601
    4. 福建师范大学数学与信息学院,福州,350007
  • 收稿日期:2019-07-17 出版日期:2019-11-30 发布日期:2019-11-29
  • 通讯作者: 夏巽鹏 E-mail:xxp1074436434@outlook.sg
  • 基金资助:
    安徽省自然科学基金(1708085MA19);安徽省级教学研究重点项目(2016jyxm0207);安徽省高校优秀青年支持计划(gxyq2017024)

Blending recommendation algorithm based ontensor decompositions and deep learning

Jiajing Zhang1,Xunpeng Xia2(),Jinlan Chen3,Youcong Ni4   

  1. 1. School of Mathematics & Physics,Anhui Jianzhu University,Hefei,230601,China
    2. School of Electronics and Information,Anhui Jianzhu University,Hefei,230601,China
    3. School of Mechanical and Electrical Engineering,Anhui Jianzhu University,Hefei,230601,China
    4. School of Math and Information Science,Fujian Normal University,Fuzhou,350007,China
  • Received:2019-07-17 Online:2019-11-30 Published:2019-11-29
  • Contact: Xunpeng Xia E-mail:xxp1074436434@outlook.sg

摘要:

张量分解和深度学习已被应用于推荐系统,并取得了较好的效果.张量分解较好地从用户对推荐对象评分中提取用户、推荐对象以及其他影响因素的隐性的特征,将这些特征进行匹配,给出推荐策略,但这种方法忽略了用户、推荐对象以及其他影响因素现有辅助数据信息中的显性特征.深度学习是从辅助信息中提取用户、推荐对象以及其他影响因素的特征,并进行匹配给出推荐策略,却忽略了用户评分数据中用户、推荐对象以及其他影响因素的隐性特征.将张量分解和深度学习两种推荐方法相融合,提出一种基于张量分解和深度学习的混合推荐算法.使用张量分解算法和深度学习分别从三阶用户评分数据和多源异构辅助信息中提取用户特征和推荐对象特征,并将它们匹配得出用户对推荐对象的需求或喜爱的预测评分,再将两种算法的预测评分进行融合给出最终综合评分,从而提高个性化推荐的精准度.对比实验证明混合推荐算法与传统的协同过滤算法相比误差降低了34.0%.

关键词: 混合推荐算法, 张量分解, 深度学习, 辅助数据, 评分数据

Abstract:

The tensor decomposition and deep learning have been applied to the recommendation systems and better results have become true. The tensor decomposition algorithm better extracts the hidden features of the users,recommended objects and other influencing factors from the user rating data,and the features are matched each other to give recommendation strategies. But the algorithm ignores the features in the auxiliary data information of the user,recommended objects and other influencing factors. Deep learning extracts the features of users,recommended objects and other influencing factors from auxiliary information,and matches them to give recommendation strategies,but ignores the implicit characteristics of users,recommended objects and other influencing factors in user rating data. A blending recommendation algorithm based on tensor decomposition and deep learning is proposed,which blends the two recommendation methods of tensor decomposition and deep learning. Tensor decomposition algorithm and deep learning are used to extract user features and recommendation object features from third?order user rating data and multi?source heterogeneous auxiliary information respectively,and then match them to obtain prediction ratings of user's demand or preference for recommendation objects,and the prediction ratings from the two algorithms are blended to give the final comprehensive ratings. The blending recommendation algorithm will improve the accuracy of personalized recommendation. Compared with traditional collaborative filtering algorithm,the error of blending recommendation algorithm is reduced by 34.0%.

Key words: blending recommendation algorithm, tensor decomposition, deep learning, auxiliary data, rating data

中图分类号: 

  • TP393

图1

张量分解和深度学习混合推荐算法示意图"

图2

深度学习部分流程图"

图3

文本卷积部分具体流程示意图"

图4

逻辑斯蒂回归示意图"

图5

模型训练损失值情况"

表1

Movielens?1M数据集上的电影推荐结果"

Precision NDCG MRR MAP
线性回归 0.2624 0.2735 0.4463 0.3902
逻辑斯蒂回归 0.3086 0.2978 0.4645 0.4132

表2

Netflix?3M数据集上的电影推荐结果"

Precision NDCG MRR MAP
线性回归 0.2832 0.2735 0.4257 0.3904
逻辑斯蒂回归 0.2905 0.2879 0.4403 0.4103

图6

单一推荐模型和混合推荐模型损失值对比图"

图7

主流推荐算法和张量深度混合模型误差比较"

1 Ramondt S , Ramírez A S . Assessing the impact of the public nutrition information environment:adapting the cancer information overload scale to measure diet information overload. Patient Education and Counseling,2018,102(1):37-42.
2 黄丹 . 基于张量分解的推荐算法研究. 硕士学位论文. 北京:北京交通大学,2016.(Hang D. Research on tensor factorization?based recommendation algorithm. Master Dissertation. Beijing:Beijing Jiaotong Universty,2016.)
3 刘凯鹏,方滨兴 . 一种基于社会性标注的网页排序算法. 计算机学报,2010,33(6):1014-1023.(Liu K P,Fang B X. A novel page ranking algorithm based on social annotations. Chinese Journal of Computers,2010,33(6):1014-1023.)
4 Zhang Z K , Liu C . Hypergraph model of social tagging network. Journal of Statistical Mechanics Theory and Experiment,2010,doi:10.1088/1742?5468/2010/10/P10005 .
doi: 10.1088/1742?5468/2010/10/P10005
5 Hofmann T . Latent semantic models for collaborative filtering. ACM Transactions on Information Systems,2004,22(1):89-115.
6 Sarwar B , Karypis G , Konstan J ,et al . Item?based collaborative filtering recommendation algorithms∥Proceedings of the 10th International Conference on World Wide Web. Hong Kong,China:ACM,2001,doi:10.1145/371920.372071.
7 Guo H F , Tang R M , Ye Y M ,et al . DeepFM:a factorization?machine based neural network for CTR prediction∥Proceedings of the 26th International Joint Conference on Artificial Intelligence. Melbourne,Australia:IJCAI,2017,doi:10.24963/ijcai.2017/239.
8 Pan R , Dolog P , Xu G . KNN?based clustering for improving social recommender systems∥Cao L,Zeng Y,Symeonidis A L, et al . Agents and Data Mining Interaction. Springer Berlin Heidelberg,2012,doi:10.1007/978?3?642?36288?0_11.
9 Liao H , Mcdermott E , Senior A . Large scale deep neural network acoustic modeling with semi?supervised training data for Youtube video trans?cription∥2013 IEEE Workshop on Automatic Speech Recognition and Understanding. Olomouc,Czech Republic:IEEE,2013,doi:10.1109/ASRU.2013.6707758.
10 邓腾飞 . 个性化音乐推荐系统的研究. 硕士学位论文. 广州:华南理工大学,2018.(Deng T F. Research of personalized music recommender system. Master Dissertation. Guangzhou:South China University of Technology,2018.)
11 Symeonidis P , Nanopoulos A , Manolopoulos Y . Tag recommendations based on tensor dimen?sionality reduction∥Proceedings of the 2008 ACM Conference on Recommender Systems. Lausanne,Switzerland:ACM,2008:43-50.
12 Rendle S , Schmidt?Thieme L . Pairwise interaction tensor factorization for personalized tag recommen?dation∥Proceedings of the 3rd ACM International Conference on Web Search and Data Mining. New York,NY,USA:ACM,2010:81-90.
13 Hsieh C K , Yang L , Cui Y ,et al . Collaborative metric learning∥Proceedings of the 26th Inter?national Conference on World Wide Web. Perth,Australia:International World Wide Web Conferences Steering Committee,2017:193-201.
14 黄立威,江碧涛,吕守业 等 . 基于深度学习的推荐系统研究综述. 计算机学报,2018,41(7):1619-1647.
Huang L W , Jiang B T , Lv S Y ,et al . Survey on deep learning based recommender systems. Chinese Journal of Computers,2018,41(7):1619-1647.
15 廖志芳,李玲,刘丽敏 等 . 三部图张量分解标签推荐算法. 计算机学报,2012,35(12):2625-2632.
Liao Z F , Li L , Liu L M ,et al . A tripartite decomposition of tensor for social tagging. Chinese Journal of Computers,2012,35(12):2625-2632.
16 Ross I M . An optimal control theory for nonlinear optimization. Journal of Computational and Applied Mathematics,2019,354:39-51.
17 Hernando A , Bobadilla J , Ortega F . A non negative matrix factorization for collaborative filtering recommender systems based on a Bayesian probabilistic model. Knowledge?Based Systems,2016,97:188-202.
18 Spiegelberg J , Rusz J , Pelckmans K . Tensor decompositions for the analysis of atomic resolution electron energy loss spectra. Ultramicroscopy,2017,175:36-45.
[1] 朱伟,张帅,辛晓燕,李文飞,王骏,张建,王炜. 结合区域检测和注意力机制的胸片自动定位与识别[J]. 南京大学学报(自然科学版), 2020, 56(4): 591-600.
[2] 李康,谢宁,李旭,谭凯. 基于卷积神经网络和几何优化的统计染色体核型分析方法[J]. 南京大学学报(自然科学版), 2020, 56(1): 116-124.
[3] 韩普,刘亦卓,李晓艳. 基于深度学习和多特征融合的中文电子病历实体识别研究[J]. 南京大学学报(自然科学版), 2019, 55(6): 942-951.
[4] 钟琪,冯亚琴,王蔚. 跨语言语料库的语音情感识别对比研究[J]. 南京大学学报(自然科学版), 2019, 55(5): 765-773.
[5] 王蔚, 胡婷婷, 冯亚琴. 基于深度学习的自然与表演语音情感识别[J]. 南京大学学报(自然科学版), 2019, 55(4): 660-666.
[6] 聂秀山1,4,王舒婷1,尹义龙2,3*. 基于特征融合和曼哈顿量化的视频哈希学习方法[J]. 南京大学学报(自然科学版), 2016, 52(4): 705-.
[7] 张鹏,黄毅,阮雅端,陈启美*. 基于稀疏特征的交通流视频检测算法[J]. 南京大学学报(自然科学版), 2015, 51(2): 264-270.
[8] 曾奎1, 何丽芳2, 杨晓伟1. 基于多线性主成分分析的支持高阶张量机[J]. 南京大学学报(自然科学版), 2014, 50(2): 219-.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 邱 浩,王欣然*. 二硫化钼的电子输运与器件[J]. 南京大学学报(自然科学版), 2014, 50(3): 280 .
[2] 王学锋1,2*,徐永兵1,2*,张 荣1,2. 低维磁性耦合体系的新物性及电/光场调控进展[J]. 南京大学学报(自然科学版), 2014, 50(3): 309 .
[3] 骆乾坤*,吴剑锋2,杨运3,钱家忠1. 渗透系数空间变异程度对进化算法优化结果影响评价[J]. 南京大学学报(自然科学版), 2015, 51(1): 60 -66 .
[4] 孙大军1,2, 王永恒1,2*, 勇俊1,2. 实频数据技术在水声换能器宽带匹配中的应用[J]. 南京大学学报(自然科学版), 2015, 51(6): 1182 -1188 .
[5] 杨政予1,王新龙1
. 茅山军号声现象的进一步研究[J]. 南京大学学报(自然科学版), 2015, 51(6): 1097 -1106 .
[6] 张亚平,2,万宇1,2,聂青3,阮晓红1,2*,王子健4. 湖泊水体中氮的生物地球化学过程及其生态学意义[J]. 南京大学学报(自然科学版), 2016, 52(1): 5 -15 .
[7] 李荣富1,2,罗跃辉 ,2,曾洪玉1,2,阮晓红1,2*,刘丛强3*. 稳定同位素技术在环境水体氮的生物地球化学循环研究中的应用[J]. 南京大学学报(自然科学版), 2016, 52(1): 16 -26 .
[8] 李 婷1,张超智1,2*,沈 丹1,袁 阳1. 石墨烯和氧化石墨烯的生物体毒性研究进展[J]. 南京大学学报(自然科学版), 2016, 52(2): 235 .
[9] 涂 臻*,卢 晶 . 散射条件下小尺度扬声器阵列声聚焦算法鲁棒性研究[J]. 南京大学学报(自然科学版), 2016, 52(2): 382 .
[10] 葛 勇1,孙宏祥1,2*,袁寿其1,夏建平1,管义钧1 . 含对称三角形腔的波导管中宽带低频隔声效应[J]. 南京大学学报(自然科学版), 2016, 52(4): 619 .