基于四阶奇异值分解的推荐算法研究
|
摘要
三阶奇异值分解推荐算法可以综合考虑用户、物品标签和物品三部分信息,挖掘三者之间的潜在关系进行推荐,然而该方法并没有引入其他方面的有效信息,如用户情感。为了考虑更多维度的信息,本文在三阶奇异值分解推荐算法的基础上,提出了一种加入用户情感信息的四阶奇异值分解推荐算法。该方法基于从评论中的emoji表情提炼出的用户情感偏好,再引入四阶张量模型,存储用户、用户情感、物品标签和物品四元组数据,应用四阶奇异值分解,从而进行个性化推荐。在某在线互联网教育的实证数据集上的实验结果表明,该方法比三阶奇异值分解推荐算法以及传统推荐算法在准确率和召回率性能指标上都有明显提升,其中进行Top-1推荐时,准确率和召回率可以达到0.513和0.339。本文的工作为移动通信端的个性化推荐提供了借鉴。
The 3 rd-order singular value decomposition recommendation algorithm can comprehensively consider the three parts of information of user, tag and item, and explore the potential relationship between the three to make recommendations. However, this method does not introduce any other effective information, such as the user's emotion. Considering more dimension information, in this paper we propose a 4 th-order singular value decomposition recommendation algorithm based on the third-order one. The method extracts user's emotional preference from the emoji expression in the commentary, introduces a 4 th-order tensor model to store user, user emotion, tag, and item quad data, and applies the 4 th-order singular value decomposition to make personalized recommendations. The experimental results on an empirical dataset of online internet education shows that the proposed method has a significant improvement in accuracy and recall performance than the third-order singular value decomposition recommendation algorithm and the traditional recommendation algorithms. In the Top-1 recommendation, the accuracy rate and recall rate of proposed method can reach 0.513 and 0.339. The work of this paper provides a reference for personalized recommendation of mobile.
The 3 rd-order singular value decomposition recommendation algorithm can comprehensively consider the three parts of information of user, tag and item, and explore the potential relationship between the three to make recommendations. However, this method does not introduce any other effective information, such as the user's emotion. Considering more dimension information, in this paper we propose a 4 th-order singular value decomposition recommendation algorithm based on the third-order one. The method extracts user's emotional preference from the emoji expression in the commentary, introduces a 4 th-order tensor model to store user, user emotion, tag, and item quad data, and applies the 4 th-order singular value decomposition to make personalized recommendations. The experimental results on an empirical dataset of online internet education shows that the proposed method has a significant improvement in accuracy and recall performance than the third-order singular value decomposition recommendation algorithm and the traditional recommendation algorithms. In the Top-1 recommendation, the accuracy rate and recall rate of proposed method can reach 0.513 and 0.339. The work of this paper provides a reference for personalized recommendation of mobile.
引文
[1]张亮,柏林森,周涛.基于跨电商行为的交叉推荐算法[J].电子科技大学学报,2013,42(1):154-160.ZHANG Liang,BAI Lin-sen,ZHOU Tao.Crossing recommendation based on multi-B2C behavior[J].Journal of University of Electronic Science&Technology of China,2013,42(1):154-160.
[2]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.ZHANG Hai-xia,LüZhen,ZHANG Chuan-ting,et al.An improved collaborative filtering recommendation algorithm with weighted heterogeneous information[J].Journal of University of Electronic Science&Technology of China,2018,47(1):112-116.
[3]丁哲,秦臻,郑文韬,等.基于移动用户浏览行为的推荐模型[J].电子科技大学学报,2017,46(6):907-912.DING Zhe,QIN Zhen,ZHENG Wen-tao,et al.Arecommendation model based on browsing behaviors of mobile users[J].Journal of University of Electronic Science&Technology of China,2017,46(6):907-912.
[4]刘建国,周涛,郭强,等.个性化推荐系统评价方法综述[J].复杂系统与复杂性科学,2009,6(3):1-10.LIU Jian-guo,ZHOU Tao,GUO Qiang,et al.Overview of the evaluated algorithms for the personal recommendation systems[J].Complex Systems&Complexity Science,2009,6(3):1-10.
[5]刘贵松,解修蕊,黄海波,等.基于最短路径信任关系的推荐项目计算方法[J].电子科技大学学报,2014,43(2):162-166.LIU Gui-song,XIE Xiu-rui,HUANG Hai-bo,et al.Fast computing method for items recommendation based on shortest-path trust relationship[J].Journal of University of Electronic Science&Technology of China,2014,43(2):162-166.
[6]ZHOU T,REN J,MEDO M,et al.Bipartite network projection and personal recommendation[J].Physical Review E,2007,76(4):046115.
[7]SARWAR B M,KARYPIS G,KONSTAN J A,et al.Application of dimensionality reduction in recommender system-a case study[C]//ACM Web kdd Workshop.[S.l.]:ACM,2000.
[8]HEYMANN P,KOUTRIKA G,GARCIA-MOLINA H.Can social bookmarking improve web search?[C]//Proceedings of the 2008 International Conference on Web Search and Data Mining.[S.l.]:ACM,2008:195-206.
[9]SYMEONIDIS P,NANOPOULOS A,MANOLOPOULOSY.Tag recommendations based on tensor dimensionality reduction[C]//Proceedings of the 2008 ACM Conference on Recommender Systems.[S.l.]:ACM,2008:43-50.
[10]SYMEONIDIS P.Clusthosvd:Item recommendation by combining semantically enhanced tag clustering with tensor HOSVD[J].IEEE Transactions on Systems Man&Cybernetics Systems,2016,46(9):1240-1251.
[11]杨君,汪会玲,艾丹祥.一种基于输出情景化的多维信息推荐新方法研究[J].情报科学,2014,32(11):126-132.YANG Jun,WANG Hui-ling,AI Dan-xiang.A new method of multi-dimensional information recommendation based on contextual post-filtering[J].Information Science,2014,32(11):126-132.
[12]SAKAGUCHI T,AKAHO Y,OKADA K,et al.Recommendation system with multi-dimensional and parallel-case four-term analogy[C]//IEEE International Conference on Systems,Man,and Cybernetics.[S.l.]:IEEE,2011:3137-3143.
[13]ADOMAVICIUS G,SANKARANARAYANAN R,SEN S,et al.Incorporating contextual information in recommender systems using a multidimensional approach[J].ACMTransactions on Information Systems,2005,23(1):103-145.
[14]CHO S,LEE M,JANG C,et al.Multidimensional filtering approach based on contextual information[C]//International Conference on Hybrid Information Technology.[S.l.]:IEEE,2006:497-504.
[15]王冠楠,陈端兵,傅彦.新闻推荐的多维兴趣模型与传播分析[J].计算机科学,2013,40(11):126-130.WANG Guan-nan,CHEN Duan-bing,FU Yan,et al.Analysis of news diffusion in recommender systems based on multidimensional tastes[J].Computer Science,2013,40(11):126-130.
[16]周娇,霍欢.社交网络服务中的多维空间视频推荐算法[J].计算机工程,2015,41(1):245-250.ZHOU Jiao,HUO Huan.Multi-dimensional space video recommendation algorithm in social network service[J].Computer Engineering,2015,41(1):245-250.
[17]YU L,LIU C,ZHANG Z K.Multi-linear interactive matrix factorization[J].Knowledge-Based Systems,2015,85:307-315.
[18]ZHANG Z K,ZHOU T,ZHANG Y C.Personalized recommendation via integrated diffusion on user-item-tag tripartite graphs[J].Physica A Statistical Mechanics&Its Applications,2009,389(1):179-186.
[19]YU L,HUANG J,ZHOU G,et al.TIIREC:A tensor approach for tag-driven item recommendation with sparse user generated content[J].Information Sciences,2017,411:122-135.
[20]LüL,MEDO M,CHI H Y,et al.Recommender systems[J].Physics Reports,2012,519(1):1-49.
[21]YUAN Q,CONG G,MA Z,et al.Time-aware point-of-interest recommendation[C]//International ACMSIGIR Conference on Research and Development in Information Retrieval.[S.l.]:ACM,2013:363-372.
[22]DING Y,LI X.Time weight collaborative filtering[C]//ACM International Conference on Information and Knowledge Management.[S.l.]:ACM,2005:485-492.
[23]李小勇,桂小林.可信网络中基于多维决策属性的信任量化模型[J].计算机学报,2009,32(3):405-416.LI Xiao-yong,GUI Xiao-ling.Trust quantitative model with multiple decision factors in trusted network[J].Chinese Journal of Computers,2009,32(3):405-416.
[24]ALKIRE S,FOSTER J.Counting and multidimensional poverty measurement[J].Journal of Public Economics,2012,95(7):476-487.
[25]裘炜毅,杨东援.基于多维分析技术的公路交通事故分析方法研究[J].交通信息与安全,2000,18(4):4-9.QIU Wei-yi,YANG Dong-yuan.Analysis of road traffic accident based on multi-dimensional analysis[J].Computer&Communications,2000,18(4):4-9.
[26]J?SCHKE R,MARINHO L,HOTHO A,et al.Tag recommendations in folksonomies[M]//Knowledge Discovery in Databases:PKDD 2007.Berlin Heidelberg:Springer,2007:506-514.
[27]BATAGELJ V,ZAVER?NIK M.Generalized cores[EB/OL].[2018-01-07].http://arXiv preprint cs/0202039,2002.
[28]KRIPPENDORFF K.Agreement and information in the reliability of coding[J].Communication Methods and Measures,2011,5(2):93-112.
[29]项亮.推荐系统实践[M].北京:人民邮电出版社,2012.XIANG Liang.Practical application of recommendation[M].Beijing:Posts&Telecom Press,2012.
[2]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.ZHANG Hai-xia,LüZhen,ZHANG Chuan-ting,et al.An improved collaborative filtering recommendation algorithm with weighted heterogeneous information[J].Journal of University of Electronic Science&Technology of China,2018,47(1):112-116.
[3]丁哲,秦臻,郑文韬,等.基于移动用户浏览行为的推荐模型[J].电子科技大学学报,2017,46(6):907-912.DING Zhe,QIN Zhen,ZHENG Wen-tao,et al.Arecommendation model based on browsing behaviors of mobile users[J].Journal of University of Electronic Science&Technology of China,2017,46(6):907-912.
[4]刘建国,周涛,郭强,等.个性化推荐系统评价方法综述[J].复杂系统与复杂性科学,2009,6(3):1-10.LIU Jian-guo,ZHOU Tao,GUO Qiang,et al.Overview of the evaluated algorithms for the personal recommendation systems[J].Complex Systems&Complexity Science,2009,6(3):1-10.
[5]刘贵松,解修蕊,黄海波,等.基于最短路径信任关系的推荐项目计算方法[J].电子科技大学学报,2014,43(2):162-166.LIU Gui-song,XIE Xiu-rui,HUANG Hai-bo,et al.Fast computing method for items recommendation based on shortest-path trust relationship[J].Journal of University of Electronic Science&Technology of China,2014,43(2):162-166.
[6]ZHOU T,REN J,MEDO M,et al.Bipartite network projection and personal recommendation[J].Physical Review E,2007,76(4):046115.
[7]SARWAR B M,KARYPIS G,KONSTAN J A,et al.Application of dimensionality reduction in recommender system-a case study[C]//ACM Web kdd Workshop.[S.l.]:ACM,2000.
[8]HEYMANN P,KOUTRIKA G,GARCIA-MOLINA H.Can social bookmarking improve web search?[C]//Proceedings of the 2008 International Conference on Web Search and Data Mining.[S.l.]:ACM,2008:195-206.
[9]SYMEONIDIS P,NANOPOULOS A,MANOLOPOULOSY.Tag recommendations based on tensor dimensionality reduction[C]//Proceedings of the 2008 ACM Conference on Recommender Systems.[S.l.]:ACM,2008:43-50.
[10]SYMEONIDIS P.Clusthosvd:Item recommendation by combining semantically enhanced tag clustering with tensor HOSVD[J].IEEE Transactions on Systems Man&Cybernetics Systems,2016,46(9):1240-1251.
[11]杨君,汪会玲,艾丹祥.一种基于输出情景化的多维信息推荐新方法研究[J].情报科学,2014,32(11):126-132.YANG Jun,WANG Hui-ling,AI Dan-xiang.A new method of multi-dimensional information recommendation based on contextual post-filtering[J].Information Science,2014,32(11):126-132.
[12]SAKAGUCHI T,AKAHO Y,OKADA K,et al.Recommendation system with multi-dimensional and parallel-case four-term analogy[C]//IEEE International Conference on Systems,Man,and Cybernetics.[S.l.]:IEEE,2011:3137-3143.
[13]ADOMAVICIUS G,SANKARANARAYANAN R,SEN S,et al.Incorporating contextual information in recommender systems using a multidimensional approach[J].ACMTransactions on Information Systems,2005,23(1):103-145.
[14]CHO S,LEE M,JANG C,et al.Multidimensional filtering approach based on contextual information[C]//International Conference on Hybrid Information Technology.[S.l.]:IEEE,2006:497-504.
[15]王冠楠,陈端兵,傅彦.新闻推荐的多维兴趣模型与传播分析[J].计算机科学,2013,40(11):126-130.WANG Guan-nan,CHEN Duan-bing,FU Yan,et al.Analysis of news diffusion in recommender systems based on multidimensional tastes[J].Computer Science,2013,40(11):126-130.
[16]周娇,霍欢.社交网络服务中的多维空间视频推荐算法[J].计算机工程,2015,41(1):245-250.ZHOU Jiao,HUO Huan.Multi-dimensional space video recommendation algorithm in social network service[J].Computer Engineering,2015,41(1):245-250.
[17]YU L,LIU C,ZHANG Z K.Multi-linear interactive matrix factorization[J].Knowledge-Based Systems,2015,85:307-315.
[18]ZHANG Z K,ZHOU T,ZHANG Y C.Personalized recommendation via integrated diffusion on user-item-tag tripartite graphs[J].Physica A Statistical Mechanics&Its Applications,2009,389(1):179-186.
[19]YU L,HUANG J,ZHOU G,et al.TIIREC:A tensor approach for tag-driven item recommendation with sparse user generated content[J].Information Sciences,2017,411:122-135.
[20]LüL,MEDO M,CHI H Y,et al.Recommender systems[J].Physics Reports,2012,519(1):1-49.
[21]YUAN Q,CONG G,MA Z,et al.Time-aware point-of-interest recommendation[C]//International ACMSIGIR Conference on Research and Development in Information Retrieval.[S.l.]:ACM,2013:363-372.
[22]DING Y,LI X.Time weight collaborative filtering[C]//ACM International Conference on Information and Knowledge Management.[S.l.]:ACM,2005:485-492.
[23]李小勇,桂小林.可信网络中基于多维决策属性的信任量化模型[J].计算机学报,2009,32(3):405-416.LI Xiao-yong,GUI Xiao-ling.Trust quantitative model with multiple decision factors in trusted network[J].Chinese Journal of Computers,2009,32(3):405-416.
[24]ALKIRE S,FOSTER J.Counting and multidimensional poverty measurement[J].Journal of Public Economics,2012,95(7):476-487.
[25]裘炜毅,杨东援.基于多维分析技术的公路交通事故分析方法研究[J].交通信息与安全,2000,18(4):4-9.QIU Wei-yi,YANG Dong-yuan.Analysis of road traffic accident based on multi-dimensional analysis[J].Computer&Communications,2000,18(4):4-9.
[26]J?SCHKE R,MARINHO L,HOTHO A,et al.Tag recommendations in folksonomies[M]//Knowledge Discovery in Databases:PKDD 2007.Berlin Heidelberg:Springer,2007:506-514.
[27]BATAGELJ V,ZAVER?NIK M.Generalized cores[EB/OL].[2018-01-07].http://arXiv preprint cs/0202039,2002.
[28]KRIPPENDORFF K.Agreement and information in the reliability of coding[J].Communication Methods and Measures,2011,5(2):93-112.
[29]项亮.推荐系统实践[M].北京:人民邮电出版社,2012.XIANG Liang.Practical application of recommendation[M].Beijing:Posts&Telecom Press,2012.