Please wait a minute...
Advanced Search
数据分析与知识发现  2020, Vol. 4 Issue (12): 76-84     https://doi.org/10.11925/infotech.2096-3467.2020.0327
     研究论文 本期目录 | 过刊浏览 | 高级检索 |
融合加权异构信息网络的矩阵分解推荐算法*
王根生1,2(),潘方正1
1江西财经大学人文学院 南昌 330013
2江西财经大学国际经贸学院 南昌 330013
Matrix Factorization Algorithm with Weighted Heterogeneous Information Network
Wang Gensheng1,2(),Pan Fangzheng1
1School of Humanities, Jiangxi University of Finance and Economics, Nanchang 330013, China
2School of International Trade and Economics, Jiangxi University of Finance and Economics,Nanchang 330013, China
全文: PDF (1236 KB)   HTML ( 13
输出: BibTeX | EndNote (RIS)      
摘要 

【目的】 在矩阵分解推荐算法中融入加权异构信息网络中的知识,提高推荐质量。【方法】 构造推荐主体的异构信息网络,通过改进的tanh函数计算相关连接权重;在网络中挑选相关元路径,通过信息增益计算不同元路径的权重;计算不同元路径下的用户兴趣相似度,再结合元路径权重得出用户相似度矩阵;将用户相似度矩阵融合到矩阵分解推荐算法的目标函数中。【结果】 在Hetrec2011-MovieLens-2k数据集上的实验结果表明,本文算法相比于传统矩阵分解推荐算法(FunkSVD)在准确率、召回率、覆盖率指标上分别提高了4.4%、5.4%、4.6%,均方根误差下降了0.06。【局限】 面对海量数据,矩阵分解的效率低下,并且没有考虑用户兴趣漂移问题。【结论】 本文算法提高了矩阵分解推荐算法的推荐质量。

服务
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章
王根生
潘方正
关键词 推荐算法矩阵分解异构信息网络元路径信息增益    
Abstract

[Objective] This paper integrates the knowledge of weighted heterogeneous information network to the matrix decomposition algorithm, aiming to improve the quality of recommendation. [Methods] First, we constructed a heterogeneous information network, and calculated the weight of connection with the improved tanh function. Then, we chose the meta paths from the network and computed their weights based on information gains. Third, we decided the similarity of user interests to create a matrix, and integrated the matrix with our algorithm. [Results] We examined the proposed algorithm with the Hetrec2011-MovieLens-2k dataset. Compared with the traditional FunkSVD algorithm, the precision, recall and coverage of our algorithm increased by 4.4%, 5.4%, and 4.6%, while its root mean square error reduced by 0.06. [Limitations] The matrix decomposition algorithm could not process massive data efficiently, and we did not investigate the drifting issues of user interests. [Conclusions] The proposed algorithm could effectively generate recommendation results.

Key wordsRecommendation Algorithms    Matrix Factorization    Heterogeneous Information Network    Meta Path    Information Gain
收稿日期: 2020-04-17      出版日期: 2020-12-25
ZTFLH:  TP393  
基金资助:*国家自然科学基金项目“社交媒体健康信息可信度评估及偏好推荐研究”(72061015);国家自然科学基金项目“自媒体环境下医患关系突发事件网络舆情演化与危机预警研究”(71461012);江西省高校人文社会科学项目“融媒体环境下网络社会舆论生态系统的空间特征及管理策略研究”(GL19110)
通讯作者: 王根生     E-mail: wgs74@126.com
引用本文:   
王根生,潘方正. 融合加权异构信息网络的矩阵分解推荐算法*[J]. 数据分析与知识发现, 2020, 4(12): 76-84.
Wang Gensheng,Pan Fangzheng. Matrix Factorization Algorithm with Weighted Heterogeneous Information Network. Data Analysis and Knowledge Discovery, 2020, 4(12): 76-84.
链接本文:  
https://manu44.magtech.com.cn/Jwk_infotech_wk3/CN/10.11925/infotech.2096-3467.2020.0327      或      https://manu44.magtech.com.cn/Jwk_infotech_wk3/CN/Y2020/V4/I12/76
Fig.1  异构信息网络实例
Fig.2  网络模式示例
Fig.3  算法流程
Fig.4  FU_w函数
元路径 语义信息
P1=User-Film-User 两用户评价过同一部电影
P2=User-Film-Genre-Film-User 两用户评价过属于同一类型下的两部电影
P3=User-Film-Actor-Film-User 两用户评价过属于同一演员主演的两部电影
P4=User-Film-Director-Film-User 两用户评价过属于同一导演执导的两部电影
P5=User-Film-Country-Film-User 两用户评价过属于同一国家上映的两部电影
P6=User-Film-Screenwriter-Film-User 两用户评价过属于同一编剧编制的两部电影
Table 1  电影推荐异构信息中典型的元路径及其语义
实体类型 表示 数目
用户 User 2 113
电影 Film 10 197
演员 Actor 95 321
导演 Director 4 060
国家 Country 72
体裁 Genre 20
评分 Rating 855 598
Table 2  不同节点的详细信息
推荐算法 用户喜爱 用户不喜爱
推荐 TP FP
未推荐 FN TN
Table 3  混合矩阵
参数名称 默认参数值
正则化参数λ2 10-3
融合系数λ1 1.5
梯度下降学习率α 10-2
梯度下降迭代次数 300
用户和电影特征维度d 20
Table 4  实验参数设置
Fig.5  不同用户和电影特征维度下的实验结果
λ1 Precision/% Recall/% Coverage/% RMSE
0.5 26.1 9.4 38.4 0.81
1.0 27.5 10.3 39.2 0.77
1.5 28.4 11.5 40.1 0.75
2.0 27.6 11.2 39.5 0.78
Table 5  不同λ1下的实验对比结果
方式 Precision/% Recall/% Coverage/% RMSE
tanh改进方式 28.4 11.5 40.1 0.75
传统方式 27.6 10.6 39.2 0.77
Table 6  tanh改进方式与直接基于原始评分方式实验结果
算法 Precision/% Recall/% Coverage/% RMSE
本文算法 28.4 11.5 40.1 0.75
FunkSVD 24.0 6.1 35.5 0.81
HIN_UCF 26.6 9.4 38.4 0.78
Table 7  不同矩阵分解推荐算法实验结果
[1] 冷亚军, 陆青, 梁昌勇 . 协同过滤推荐技术综述[J]. 模式识别与人工智能, 2014,27(8):720-734.
[1] ( Leng Yajun, Lu Qing, Liang Changyong . Survey of Recommendation Based on Collaborative Filtering[J]. Pattern Recognition and Artificial Intelligence, 2014,27(8):720-734.)
[2] 焦富森, 李树青 . 基于物品质量和用户评分修正的协同过滤推荐算法[J]. 数据分析与知识发现, 2019,3(8):62-67.
[2] ( Jiao Fusen, Li Shuqing . Collaborative Filtering Recommendation Based on Item Quality and User Ratings[J]. Data Analysis and Knowledge Discovery, 2019,3(8):62-67.)
[3] Zhang H Y, Ganchev I, Nikolov N S, et al. A Trust-enriched Approach for Item-based Collaborative Filtering Recommendations [C]//Proceedings of the 12th International Conference on Intelligent Computer Communication and Processing (ICCP). 2016: 65-68.
[4] 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[J]. 计算机学报, 2018,41(7):1619-1647.
[4] ( Huang Liwei, Jiang Bitao, Lv Shouye , et al. Survey on Deep Learning Based Recommender Systems[J]. Chinese Journal of Computers, 2018,41(7):1619-1647.)
[5] Forsati R, Mahdavi M, Shamsfard M , et al. Matrix Factorization with Explicit Trust and Distrust Side Information for Improved Social Recommendation[J]. ACM Transactions on Information Systems, 2014, 32(4): Article No. 17.
[6] Yu Y H, Wang C, Wang H , et al. Attributes Coupling Based Matrix Factorization for Item Recommendation[J]. Applied Intelligence, 2017,46(3):521-533.
[7] Chen S L, Peng Y X . Matrix Factorization for Recommendation with Explicit and Implicit Feedback[J]. Knowledge-Based Systems, 2018,158:109-117.
[8] 李昆仑, 郭昌隆, 关立伟 . 一种融合近邻用户影响力的矩阵分解推荐算法[J]. 小型微型计算机系统, 2018,39(1):37-41.
[8] ( Li Kunlun, Guo Changlong, Guan Liwei . Matrix Factorization Recommendation Algorithm Based on the Influence of Nearest Neighbor Users[J]. Journal of Chinese Computer Systems, 2018,39(1):37-41.)
[9] 余永红, 高阳, 王皓 , 等. 融合用户社会地位和矩阵分解的推荐算法[J]. 计算机研究与发展, 2018,55(1):113-124.
[9] ( Yu Yonghong, Gao Yang, Wang Hao , et al. Integrating User Social Status and Matrix Factorization for Item Recommendation[J]. Journal of Computer Research and Development, 2018,55(1):113-124.)
[10] 文凯, 朱传亮 . 融合社交网络和兴趣的正则化矩阵分解推荐模型[J]. 计算机应用, 2018,38(9):2523-2528.
[10] ( Wen Kai, Zhu Chuanliang . Regularized Matrix Decomposition Recommendation Model Integrating Social Networks and Interest Correlation[J]. Journal of Computer Applications, 2018,38(9):2523-2528.)
[11] Chen Y N, Liu R F, Xu W R. Movie Recommendation in Heterogeneous Information Networks [C]//Proceedings of 2016 IEEE Information Technology, Networking, Electronic and Automation Control Conference. 2016: 637-640.
[12] Ma X, Zhang Y, Zeng J F . Newly Published Scientific Papers Recommendation in Heterogeneous Information Networks[J]. Mobile Networks & Applications, 2019,24(1):69-79.
[13] Gupta M, Kumar P . Recommendation Generation Using Personalized Weight of Meta-paths in Heterogeneous Information Networks[J]. European Journal of Operational Research, 2020,284(2):660-674.
doi: 10.1016/j.ejor.2020.01.010
[14] 石磊, 丁鑫, 陶永才 , 等. 一种检测兴趣漂移的元路径推荐模型[J]. 小型微型计算机系统, 2019,40(3):612-617.
[14] ( Shi Lei, Ding Xin, Tao Yongcai , et al. Meta Path Recommendation Model with Interest Drift Detecting[J]. Journal of Chinese Computer Systems, 2019,40(3):612-617.)
[15] 王旭, 庞巍, 王喆 . 异构信息网络中基于元结构的协同过滤算法[J]. 计算机科学, 2019,46(6A):397-401, 416.
[15] ( Wang Xu, Pang Wei, Wang Zhe . MetaStruct-CF:A Meta Structure Based Collaborative Filtering Algorithm in Heterogeneous Information Networks[J]. Computer Science, 2019,46(6A):397-401, 416.)
[16] 张海霞, 吕振, 张传亭 , 等. 一种引入加权异构信息的改进协同过滤推荐算法[J]. 电子科技大学学报, 2018,47(1):112-116, 152.
[16] ( Zhang Haixia, Lv Zhen, Zhang Chuanting , et al. An Improved Collaborative Filtering Recommendation Algorithm with Weighted Heterogeneous Information[J]. Journal of University of Electronic Science and Technology of China, 2018,47(1):112-116, 152.)
[17] 赵传, 张凯涵, 梁吉业 . 非对称的异质信息网络推荐算法[J]. 计算机科学与探索, 2020,14(6):939-946.
[17] ( Zhao Chuan, Zhang Kaihan, Liang Jiye . Asymmetric Recommendation Algorithm in Heterogeneous Information Network[J]. Journal of Frontiers of Computer Science and Technology, 2020,14(6):939-946.)
[18] Koren Y, Bell R, Volinsky C . Matrix Factorization Techniques for Recommender Systems[J]. Computer, 2009,42(8):30-37.
[19] 李改, 李磊 . 基于矩阵分解的协同过滤算法[J]. 计算机工程与应用, 2011,47(30):4-7.
[19] ( Li Gai, Li Lei . Collaborative Filtering Algorithm Based on Matrix Decomposition[J]. Computer Engineering and Applications, 2011,47(30):4-7.)
[20] Shi C, Li Y T, Zhang J W , et al. A Survey of Heterogeneous Information Network Analysis[J]. IEEE Transactions on Knowledge & Data Engineering, 2015,29(1):17-37.
[21] 王根生, 黄学坚, 吴小芳 , 等. 基于改进信息增益特征选择法的SVM中文情感分类算法[J]. 成都理工大学学报(自然科学版), 2019,46(1):105-110.
[21] ( Wang Gensheng, Huang Xuejian, Wu Xiaofang , et al. Algorithm of Chinese Sentiment Classification of SVM Based on Optimization Information Gain Feature Selection Method[J]. Journal of Chengdu University of Technology(Science & Technology Edition) , 2019,46(1):105-110.)
[22] Pathak A, Chakrabarti S, Gupta M. Index Design for Dynamic Personalized PageRank [C]//Proceedings of the 24th International Conference on Data Engineering. 2008: 1489-1491.
[23] Jeh G, Widom J. SimRank: A Measure of Structural-context Similarity [C]//Proceedings of the 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2002: 538-543.
[24] Sun Y Z, Han J W, Yan X F, et al. PathSim: Meta Path-Based Top-K Similarity Search in Heterogeneous Information Networks [C]//Proceedings of 2011 International Conference on Very Large Data Bases (VLDB’11). 2011: 992-1003.
[1] 王勤洁, 秦春秀, 马续补, 刘怀亮, 徐存真. 基于作者偏好和异构信息网络的科技文献推荐方法研究*[J]. 数据分析与知识发现, 2021, 5(8): 54-64.
[2] 马莹雪,甘明鑫,肖克峻. 融合标签和内容信息的矩阵分解推荐方法*[J]. 数据分析与知识发现, 2021, 5(5): 71-82.
[3] 杨辰, 陈晓虹, 王楚涵, 刘婷婷. 基于用户细粒度属性偏好聚类的推荐策略*[J]. 数据分析与知识发现, 2021, 5(10): 94-102.
[4] 张纯金,郭盛辉,纪淑娟,杨伟,伊磊. 基于多属性评分隐表征学习的群组推荐算法*[J]. 数据分析与知识发现, 2020, 4(12): 120-135.
[5] 温彦,马立健,曾庆田,郭文艳. 基于地理信息偏好修正和社交关系偏好隐式分析的POI推荐 *[J]. 数据分析与知识发现, 2019, 3(8): 30-39.
[6] 丁勇,程璐,蒋翠清. 基于二部图的P2P网络借贷投资组合决策方法 *[J]. 数据分析与知识发现, 2019, 3(12): 76-83.
[7] 侯君, 刘魁, 李千目. 基于ESSVM的分类推荐*[J]. 数据分析与知识发现, 2018, 2(3): 9-21.
[8] 施晓华, 卢宏涛. 基于矩阵分解学习的科学合作网络社区发现研究*[J]. 数据分析与知识发现, 2017, 1(9): 49-56.
[9] 王永,邓江洲,邓永恒,张璞. 基于项目概率分布的协同过滤推荐算法*[J]. 现代图书情报技术, 2016, 32(6): 73-79.
[10] 陈东沂,周子程,蒋盛益,王连喜,吴佳林. 面向企业微博的客户细分框架*[J]. 现代图书情报技术, 2016, 32(2): 43-51.
[11] 王伟军, 宋梅青. 一种面向用户偏好定向挖掘的协同过滤个性化推荐算法[J]. 现代图书情报技术, 2014, 30(6): 25-32.
[12] 原福永, 蔡红蕾. 一种在信任网络中随机游走的推荐算法[J]. 现代图书情报技术, 2014, 30(10): 70-75.
[13] 陈涛,宋妍,谢阳群. 改进的信息增益特征选择方法在文本聚类中的应用*[J]. 现代图书情报技术, 2004, 20(12): 7-9.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
版权所有 © 2015 《数据分析与知识发现》编辑部
地址:北京市海淀区中关村北四环西路33号 邮编:100190
电话/传真:(010)82626611-6626,82624938
E-mail:jishu@mail.las.ac.cn