Truser:一种基于可信用户的服务推荐方法
|
摘要
在服务推荐过程中,为排除不可信用户信息带来的干扰,确保推荐结果的精准性,该文从用户聚类的角度,通过两阶段的ISODATA聚类,将离群用户视为不可信用户进行过滤,再基于得到的可信用户提出一种改进的服务推荐方法.最后,在两个公开数据集Last.FM和Delicious上进行了实证分析.结果表明,该文所提方法在两个数据集上的推荐精度相较于已有基准方法分别提高16.1%和4.5%,且发现当第一阶段聚类的预期聚类中心为6时,推荐效果最好;同时,在推荐过程中为目标用户返回Top-5个可信用户,且向其推荐这5个用户中至少有70%的人关注过的服务最为适宜.因此,围绕可信用户的数据进行推荐,能有效地提高服务推荐的质量.
As a very important topic in the field of service computing,service recommendation has been paid much attention by researchers.To improve the service recommendation quality and ensure user's experience of services,various methods have been proposed successively.However,most existing methods mainly focus on how to improve the accuracy of recommendation models by introducing richer information or advanced modeling techniques,and there is a general lack of discussion on users' trustworthiness from the data quality point of view.In practice,untrusted users are common for a variety of reasons,such as fake comments.Data generated by untrusted users is often misleading and unhelpful for recommendations.Therefore,it is very necessary to eliminate noise from untrusted users before service recommendation;otherwise,the quality of recommendations will always be affected regardless of how the model is optimized.To achieve this,we identify untrusted users based on their abnormal labeling behavior compared to the public and attempt to filter out these outliers before service recommendation,from the perspective of a two-stage ISODATA(Iterative Self-Organizing Data Analysis Technique Algorithm) clustering,and then propose a novel approach to service recommendation based on the resulting trusted users,named as Truser.Compared with the conventional K-means clustering algorithm,ISODATA is more flexible and increases the operation of"merging"and "splitting"to adjust the clustering number.In other words,the clustering center K value of ISODATA can be adjusted dynamically according to the actual situation.In this paper,we first perform ISODATA clustering for the concerned users of each service to label its candidate untrusted users.Then we get the number of times each user is marked as candidate untrusted,and the number of times any two users are clustered into the same group.According to the information obtained above,we further utilize ISODATA to cluster users,so as to filter out untrusted users eventually.Finally,we return Top-k similar users for the target user in the set of trusted users,and implement service recommendation based on the selection of similar trusted users.To verify the feasibility of our approach,an empirical study is conducted on two public datasets:Last.FM and Delicious.We compare the proposed method with other state-of-the-art recommendation methods with the general performance metric:Root Mean Square Error(RMSE).Experimental results show that,compared with the existing methods,the accuracies of recommendation of our approach achieved on the two datasets are improved by 16.1% and 4.5%,respectively.By learning the parameters of the proposed algorithm,we find that the recommendation result is the best when the expected clusters kis set to six during the first-stage clustering.Meanwhile,the experimental results also suggest that it is most suitable for the target user to return the Top-5 most similar trusted users,and to recommend services followed by at least 70% of those similar users.In general,the results roughly coincide with the practice.Therefore,it is reasonable to conclude that the quality of service recommendations can be improved by filtering out untrusted users.
As a very important topic in the field of service computing,service recommendation has been paid much attention by researchers.To improve the service recommendation quality and ensure user's experience of services,various methods have been proposed successively.However,most existing methods mainly focus on how to improve the accuracy of recommendation models by introducing richer information or advanced modeling techniques,and there is a general lack of discussion on users' trustworthiness from the data quality point of view.In practice,untrusted users are common for a variety of reasons,such as fake comments.Data generated by untrusted users is often misleading and unhelpful for recommendations.Therefore,it is very necessary to eliminate noise from untrusted users before service recommendation;otherwise,the quality of recommendations will always be affected regardless of how the model is optimized.To achieve this,we identify untrusted users based on their abnormal labeling behavior compared to the public and attempt to filter out these outliers before service recommendation,from the perspective of a two-stage ISODATA(Iterative Self-Organizing Data Analysis Technique Algorithm) clustering,and then propose a novel approach to service recommendation based on the resulting trusted users,named as Truser.Compared with the conventional K-means clustering algorithm,ISODATA is more flexible and increases the operation of"merging"and "splitting"to adjust the clustering number.In other words,the clustering center K value of ISODATA can be adjusted dynamically according to the actual situation.In this paper,we first perform ISODATA clustering for the concerned users of each service to label its candidate untrusted users.Then we get the number of times each user is marked as candidate untrusted,and the number of times any two users are clustered into the same group.According to the information obtained above,we further utilize ISODATA to cluster users,so as to filter out untrusted users eventually.Finally,we return Top-k similar users for the target user in the set of trusted users,and implement service recommendation based on the selection of similar trusted users.To verify the feasibility of our approach,an empirical study is conducted on two public datasets:Last.FM and Delicious.We compare the proposed method with other state-of-the-art recommendation methods with the general performance metric:Root Mean Square Error(RMSE).Experimental results show that,compared with the existing methods,the accuracies of recommendation of our approach achieved on the two datasets are improved by 16.1% and 4.5%,respectively.By learning the parameters of the proposed algorithm,we find that the recommendation result is the best when the expected clusters kis set to six during the first-stage clustering.Meanwhile,the experimental results also suggest that it is most suitable for the target user to return the Top-5 most similar trusted users,and to recommend services followed by at least 70% of those similar users.In general,the results roughly coincide with the practice.Therefore,it is reasonable to conclude that the quality of service recommendations can be improved by filtering out untrusted users.
引文
[1]Linden G,Smith B,York J.Amazon.com recommendations:Item-to-item collaborative filtering.IEEE Internet Computing,2003,7(1):76-80
[2]Wei S,Zheng X,Chen D,et al.A hybrid approach for movie recommendation via tags and ratings.Electronic Commerce Research&Applications,2016,18(C):83-94
[3]Lawrence E.Everything is a recommendation netflix,altgenres and the construction of taste.Knowledge Organization,2015,42(5):358-364
[4]Xia Li-Xin,Li Chong-Yang,Cheng Xiu-Feng,et al.Multi dimension recommendation model of heterogeneous network resources-A case study of douban.Library&Information Service,2017,61(3):6-13(in Chinese)(夏立新,李重阳,程秀峰等.异质网络资源多维度推荐模式研究——以豆瓣网为例.图书情报工作,2017,61(3):6-13)
[5]Benzi K,Kalofolias V,Bresson X,et al.Song recommendation with non-negative matrix factorization and graph total variation//Proceedings of the IEEE International Conference on Acoustics,Speech and Signal Processing.New Orleans,USA,2016:2439-2443
[6]Zheng Z,Ma H,Lyu M R,et al.WSRec:A collaborative filtering based Web service recommender system//Proceedings of the IEEE International Conference on Web Services.Los Angeles,USA,2009:437-444
[7]Yao L,Sheng Q Z,Ngu A H H,et al.Unified collaborative and content-based Web service recommendation.IEEETransactions on Services Computing,2015,8(3):453-466
[8]Zhang Y,Zheng Z,Lyu M R.WSPred:A time-aware personalized QoS prediction framework for Web services//Proceedings of the IEEE International Symposium on Software Reliability Engineering.Dallas,USA,2012:210-219
[9]Xie Q,Zhao S,Zheng Z,et al.Asymmetric correlation regularized matrix factorization for Web service recommendation//Proceedings of the IEEE International Conference on Web Services.San Francisco,USA,2016:204-211
[10]Zheng Z,Ma H,Lyu M R,et al.Collaborative Web service QoS prediction via neighborhood integrated matrix factorization.IEEE Transactions on Services Computing,2013,6(3):289-299
[11]Wang S,Lo D,Vasilescu B,et al.EnTagRec++:An enhanced tag recommendation system for software information sites.Empirical Software Engineering,2018,23(3):800-832
[12]Qiu W,Zheng Z,Wang X,et al.Reputation-aware QoSvalue prediction of Web services//Proceedings of the IEEEInternational Conference on Services Computing.Santa Clara,USA,2013:41-48
[13]Noorian Z,Marsh S,Fleming M.zTrust:Adaptive decentralized trust model for quality of service selection in electronic marketplaces.Computational Intelligence,2016,32(1):127-164
[14]Kim Y A.A trust prediction framework in rating-based experience sharing social networks without a Web of Trust.Information Sciences,2012,191(9):128-145
[15]Wu C,Qiu W,Zheng Z,et al.QoS prediction of Web services based on two-phase K-means clustering//Proceedings of the IEEE International Conference on Web Services.New York,USA,2015:161-168
[16]Pall G H.ISODATA:An interactive method of multivariate data analysis and pattern classification//Proceedings of the IEEE International Communication Conference.1966:116-117
[17]Shao L,Zhang J,Wei Y,et al.Personalized QoS prediction for Web services via collaborative filtering//Proceedings of the IEEE International Conference on Web Services.Salt Lake City,USA,2007:439-446
[18]Wang H,Tao Y,Yu Q,et al.Incorporating both qualitative and quantitative preferences for service recommendation.Journal of Parallel&Distributed Computing,2018,114:46-69
[19]Hu Q Y,Zhao Z L,Wang C D,et al.An item orientated recommendation algorithm from the multi-view perspective.Neurocomputing,2017,269:261-272
[20]Chen M,Ma Y,Hu B,et al.A ranking-oriented hybrid approach to QoS-aware Web service recommendation//Proceedings of the IEEE International Conference on Services Computing.New York City,USA,2015:578-585
[21]Ren L,Wang W.An SVM-based collaborative filtering approach for Top-N Web services recommendation.Future Generation Computer Systems,2017,78:531-543
[22]Deng S,Huang L,Xu G.Social network-based service recommendation with trust enhancement.Expert Systems with Applications,2014,41(18):8075-8084
[23]Cao B,Liu J,Tang M,et al.Mashup service recommendation based on user Interest and social network//Proceedings of the IEEE International Conference on Web Services.Santa Clara,USA,2013:99-106
[24]Xu J,Zheng Z,Lyu M R.Web service personalized quality of service prediction via reputation-based matrix factorization.IEEE Transactions on Reliability,2016,65(1):28-37
[25]Chen X,Liu X,Huang Z,et al.RegionKNN:A scalable hybrid collaborative filtering algorithm for personalized Web service recommendation//Proceedings of the IEEE International Conference on Web Services.Miami,USA,2010:9-16
[26]Wang S,Zhao Y,Huang L,et al.QoS prediction for service recommendations in mobile edge computing.Journal of Parallel&Distributed Computing,2017,DOI:https://doi.org/10.1016/j.jpdc.2017.09.014
[27]Xie Qi,Wu Kai-Gui,Xu Jie,et al.Personalized contextaware QoS prediction for Web services based on collaborative filtering//Proceedings of the International Conference on Advanced Data Mining and Applications.Brussels,Belgium,2010:368-375
[28]Wang H,Wang L,Yu C.Integrating trust with qualitative and quantitative preference for service selection//Proceedings of the IEEE International Conference on Services Computing.Anchorage,USA,2014:299-306
[29]Wei J.A personalized authoritative user-based recommendation for social tagging.Future Generation Computer Systems,2018,86:355-361
[30]Xu R.Survey of Clustering Algorithms.New York,USA:IEEE Press,2005
[31]Yoon K A,Kwon O S,Bae D H.An approach to outlier detection of software measurement data using the K-means clustering method//Proceedings of the International Symposium on Empirical Software Engineering and Measurement.Madrid,Spain,2007:443-445
[32]Wu J,Chen L,Feng Y,et al.Predicting quality of service for selection by neighborhood-based collaborative filtering.IEEE Transactions on Systems Man&Cybernetics Systems,2013,43(2):428-439
[33]Pan Wei-Feng,Li Bing,Shao Bo,et al.Service classification and recommendation based on software networks.Chinese Journal of Computers,2011,34(12):2355-2369(in Chinese)(潘伟丰,李兵,邵波等.基于软件网络的服务自动分类和推荐方法研究.计算机学报,2011,34(12):2355-2369)
[34]Kim D,Kim K S,Park K H,et al.A music recommendation system with a dynamic K-means clustering algorithm//Proceedings of the International Conference on Machine Learning and Applications.San Diego,USA,2008:399-403
[35]He Peng,Li Bing,Yang Xi-Hui,et al.Roster:An approach to potential peer recommendation for developers.Chinese Journal of Computers,2014,37(4):859-872(in Chinese)(何鹏,李兵,杨习辉等.Roster:一种开发者潜在同行推荐方法.计算机学报,2014,37(4):859-872)
[36]Sarwar B,Karypis G,Konstan J,et al.Item-based collaborative filtering recommendation algorithms//Proceedings of the International Conference on World Wide Web.Hyderabad,India,2001:285-295
[37]Wang C,Blei D M.Collaborative topic modeling for recommending scientific articles//Proceedings of the ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.San Diego,USA,2011:448-456
[38]Mnih A,Salakhutdinov R R.Probabilistic matrix factorization.Advances in Neural Information Processing Systems,2007,20(2):1257-1264
[39]Kang G,Liu J,Tang M,et al.AWSR:Active Web service recommendation based on usage history//Proceedings of the International Conference on Web Services.Honolulu,USA,2012:186-193
[40]Wang H,Chi X,Wang Z,et al.Extracting fine-grained service value features and distributions for accurate service recommendation//Proceedings of the IEEE International Conference on Web Services.Honolulu,USA,2017:277-284
[2]Wei S,Zheng X,Chen D,et al.A hybrid approach for movie recommendation via tags and ratings.Electronic Commerce Research&Applications,2016,18(C):83-94
[3]Lawrence E.Everything is a recommendation netflix,altgenres and the construction of taste.Knowledge Organization,2015,42(5):358-364
[4]Xia Li-Xin,Li Chong-Yang,Cheng Xiu-Feng,et al.Multi dimension recommendation model of heterogeneous network resources-A case study of douban.Library&Information Service,2017,61(3):6-13(in Chinese)(夏立新,李重阳,程秀峰等.异质网络资源多维度推荐模式研究——以豆瓣网为例.图书情报工作,2017,61(3):6-13)
[5]Benzi K,Kalofolias V,Bresson X,et al.Song recommendation with non-negative matrix factorization and graph total variation//Proceedings of the IEEE International Conference on Acoustics,Speech and Signal Processing.New Orleans,USA,2016:2439-2443
[6]Zheng Z,Ma H,Lyu M R,et al.WSRec:A collaborative filtering based Web service recommender system//Proceedings of the IEEE International Conference on Web Services.Los Angeles,USA,2009:437-444
[7]Yao L,Sheng Q Z,Ngu A H H,et al.Unified collaborative and content-based Web service recommendation.IEEETransactions on Services Computing,2015,8(3):453-466
[8]Zhang Y,Zheng Z,Lyu M R.WSPred:A time-aware personalized QoS prediction framework for Web services//Proceedings of the IEEE International Symposium on Software Reliability Engineering.Dallas,USA,2012:210-219
[9]Xie Q,Zhao S,Zheng Z,et al.Asymmetric correlation regularized matrix factorization for Web service recommendation//Proceedings of the IEEE International Conference on Web Services.San Francisco,USA,2016:204-211
[10]Zheng Z,Ma H,Lyu M R,et al.Collaborative Web service QoS prediction via neighborhood integrated matrix factorization.IEEE Transactions on Services Computing,2013,6(3):289-299
[11]Wang S,Lo D,Vasilescu B,et al.EnTagRec++:An enhanced tag recommendation system for software information sites.Empirical Software Engineering,2018,23(3):800-832
[12]Qiu W,Zheng Z,Wang X,et al.Reputation-aware QoSvalue prediction of Web services//Proceedings of the IEEEInternational Conference on Services Computing.Santa Clara,USA,2013:41-48
[13]Noorian Z,Marsh S,Fleming M.zTrust:Adaptive decentralized trust model for quality of service selection in electronic marketplaces.Computational Intelligence,2016,32(1):127-164
[14]Kim Y A.A trust prediction framework in rating-based experience sharing social networks without a Web of Trust.Information Sciences,2012,191(9):128-145
[15]Wu C,Qiu W,Zheng Z,et al.QoS prediction of Web services based on two-phase K-means clustering//Proceedings of the IEEE International Conference on Web Services.New York,USA,2015:161-168
[16]Pall G H.ISODATA:An interactive method of multivariate data analysis and pattern classification//Proceedings of the IEEE International Communication Conference.1966:116-117
[17]Shao L,Zhang J,Wei Y,et al.Personalized QoS prediction for Web services via collaborative filtering//Proceedings of the IEEE International Conference on Web Services.Salt Lake City,USA,2007:439-446
[18]Wang H,Tao Y,Yu Q,et al.Incorporating both qualitative and quantitative preferences for service recommendation.Journal of Parallel&Distributed Computing,2018,114:46-69
[19]Hu Q Y,Zhao Z L,Wang C D,et al.An item orientated recommendation algorithm from the multi-view perspective.Neurocomputing,2017,269:261-272
[20]Chen M,Ma Y,Hu B,et al.A ranking-oriented hybrid approach to QoS-aware Web service recommendation//Proceedings of the IEEE International Conference on Services Computing.New York City,USA,2015:578-585
[21]Ren L,Wang W.An SVM-based collaborative filtering approach for Top-N Web services recommendation.Future Generation Computer Systems,2017,78:531-543
[22]Deng S,Huang L,Xu G.Social network-based service recommendation with trust enhancement.Expert Systems with Applications,2014,41(18):8075-8084
[23]Cao B,Liu J,Tang M,et al.Mashup service recommendation based on user Interest and social network//Proceedings of the IEEE International Conference on Web Services.Santa Clara,USA,2013:99-106
[24]Xu J,Zheng Z,Lyu M R.Web service personalized quality of service prediction via reputation-based matrix factorization.IEEE Transactions on Reliability,2016,65(1):28-37
[25]Chen X,Liu X,Huang Z,et al.RegionKNN:A scalable hybrid collaborative filtering algorithm for personalized Web service recommendation//Proceedings of the IEEE International Conference on Web Services.Miami,USA,2010:9-16
[26]Wang S,Zhao Y,Huang L,et al.QoS prediction for service recommendations in mobile edge computing.Journal of Parallel&Distributed Computing,2017,DOI:https://doi.org/10.1016/j.jpdc.2017.09.014
[27]Xie Qi,Wu Kai-Gui,Xu Jie,et al.Personalized contextaware QoS prediction for Web services based on collaborative filtering//Proceedings of the International Conference on Advanced Data Mining and Applications.Brussels,Belgium,2010:368-375
[28]Wang H,Wang L,Yu C.Integrating trust with qualitative and quantitative preference for service selection//Proceedings of the IEEE International Conference on Services Computing.Anchorage,USA,2014:299-306
[29]Wei J.A personalized authoritative user-based recommendation for social tagging.Future Generation Computer Systems,2018,86:355-361
[30]Xu R.Survey of Clustering Algorithms.New York,USA:IEEE Press,2005
[31]Yoon K A,Kwon O S,Bae D H.An approach to outlier detection of software measurement data using the K-means clustering method//Proceedings of the International Symposium on Empirical Software Engineering and Measurement.Madrid,Spain,2007:443-445
[32]Wu J,Chen L,Feng Y,et al.Predicting quality of service for selection by neighborhood-based collaborative filtering.IEEE Transactions on Systems Man&Cybernetics Systems,2013,43(2):428-439
[33]Pan Wei-Feng,Li Bing,Shao Bo,et al.Service classification and recommendation based on software networks.Chinese Journal of Computers,2011,34(12):2355-2369(in Chinese)(潘伟丰,李兵,邵波等.基于软件网络的服务自动分类和推荐方法研究.计算机学报,2011,34(12):2355-2369)
[34]Kim D,Kim K S,Park K H,et al.A music recommendation system with a dynamic K-means clustering algorithm//Proceedings of the International Conference on Machine Learning and Applications.San Diego,USA,2008:399-403
[35]He Peng,Li Bing,Yang Xi-Hui,et al.Roster:An approach to potential peer recommendation for developers.Chinese Journal of Computers,2014,37(4):859-872(in Chinese)(何鹏,李兵,杨习辉等.Roster:一种开发者潜在同行推荐方法.计算机学报,2014,37(4):859-872)
[36]Sarwar B,Karypis G,Konstan J,et al.Item-based collaborative filtering recommendation algorithms//Proceedings of the International Conference on World Wide Web.Hyderabad,India,2001:285-295
[37]Wang C,Blei D M.Collaborative topic modeling for recommending scientific articles//Proceedings of the ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining.San Diego,USA,2011:448-456
[38]Mnih A,Salakhutdinov R R.Probabilistic matrix factorization.Advances in Neural Information Processing Systems,2007,20(2):1257-1264
[39]Kang G,Liu J,Tang M,et al.AWSR:Active Web service recommendation based on usage history//Proceedings of the International Conference on Web Services.Honolulu,USA,2012:186-193
[40]Wang H,Chi X,Wang Z,et al.Extracting fine-grained service value features and distributions for accurate service recommendation//Proceedings of the IEEE International Conference on Web Services.Honolulu,USA,2017:277-284
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |