社交网络下基于列表级排序学习的推荐算法
|
摘要
将列表级排序学习和推荐算法相结合,能够有效提高传统推荐系统返回结果的准确性。针对社交网络环境,提出一种基于列表级排序学习的推荐算法L~2R~2SN (list-wise learning to rank for recommendation for social networks)。从社交网络中挖掘出用户好友潜在的影响特征,以及物品潜在的隐性特征,融入列表级排序学习的推荐模型中,通过梯度下降方法迭代训练模型参数获得模型的最优解,将物品列表中排序较前的top-k个物品推送给用户。多组实验结果表明,L~2R~2SN算法能够有效提高推荐结果的准确性,更为有效地反映用户的偏好。
Combining list-wise learning to rank and recommendation algorithms can improve the accuracy of results of traditional recommendation systems.Based on list-wise learning to rank,an efficient recommendation algorithm L~2R~2SN(list-wise learning to rank for recommendation for social networks)on social network environments was proposed.The potential impact features of user's friends from social networks and the potential hidden features of items were extracted,and these features were integrated into the list-wise learning to rank based recommendation models.The strategy of gradient descent iterative training was used to obtain the optimal solution of the models.The top-kitems were recommended to the users.Experimental results show that the L~2R~2SN algorithm can improve the accuracy of the recommendation results effectively and reflect the user's preference more effectively.
Combining list-wise learning to rank and recommendation algorithms can improve the accuracy of results of traditional recommendation systems.Based on list-wise learning to rank,an efficient recommendation algorithm L~2R~2SN(list-wise learning to rank for recommendation for social networks)on social network environments was proposed.The potential impact features of user's friends from social networks and the potential hidden features of items were extracted,and these features were integrated into the list-wise learning to rank based recommendation models.The strategy of gradient descent iterative training was used to obtain the optimal solution of the models.The top-kitems were recommended to the users.Experimental results show that the L~2R~2SN algorithm can improve the accuracy of the recommendation results effectively and reflect the user's preference more effectively.
引文
[1]Wang F,Mack E A,Maciewjewski R.Analyzing entrepreneurial social networks with big data[J].Annals of the American Association of Geographers,2017,107(1):130-150.
[2]Lu J,Wu D,Mao M,et al.Recommender system application developments:A survey[J].Decision Support Systems,2015,74(C):12-32.
[3]Hernando A,Bobadilla J,Ortega F.A non negative matrix factorization for collaborative filtering recommender systems based on a Bayesian probabilistic model[J].Knowledge-Based Systems,2016,97(C):188-202.
[4]Wang H,Wang N,Yeung D Y.Collaborative deep learning for recommender systems[C]//Proceedings of the 21th ACMSIGKDD International Conference on Knowledge Discovery and Data Mining.ACM,2015:1235-1244.
[5]Paradarami T K,Bastian N D,Wightman J L.A hybrid recommender system using artificial neural networks[J].Expert Systems with Applications,2017,83(C):300-313.
[6]Capdevila J,Arias M,Arratia A.GeoSRS:A hybrid social recommender system for geolocated data[J].Information Systems,2016,57(1):111-128.
[7]Karatzoglou A,Baltrunas L,Shi Y.Learning to rank for recommender systems[C]//Proceedings of the 7th ACM Conference on Recommender Systems.ACM,2013:493-494.
[8]Shi Y,Larson M,Hanjalic A.List-wise learning to rank with matrix factorization for collaborative filtering[C]//Proceedings of the Fourth ACM Conference on Recommender Systems.ACM,2010:269-272.
[9]Gray-DaviesT,Holmes C C,Caron F.Scalable Bayesian nonparametric regression via a plackett-Luce model for conditional ranks[J].Electronic Journal of Statistics,2016,10(2):1807-1828.
[10]Weimer M,Karatzoglou A,Le Q V,et al.COFI RANK,Maximum margin matrix factorization for collaborative ranking[C]//International Conference on Neural Information Processing Systems.Curran Associates Inc,2008:1593-1600.
[11]Yao W,He J,Huang G,et al.SoRank:Incorporating social information into learning to rank models for recommendation[C]//Proceedings of the 23rd International Conference on World Wide Web.ACM,2014:409-410.
[12]Kumar V,Pujari A K,Sahu S K,et al.Proximal maximum margin matrix factorization for collaborative filtering[J].Pattern Recognition Letters,2017,86(1):62-67.
[2]Lu J,Wu D,Mao M,et al.Recommender system application developments:A survey[J].Decision Support Systems,2015,74(C):12-32.
[3]Hernando A,Bobadilla J,Ortega F.A non negative matrix factorization for collaborative filtering recommender systems based on a Bayesian probabilistic model[J].Knowledge-Based Systems,2016,97(C):188-202.
[4]Wang H,Wang N,Yeung D Y.Collaborative deep learning for recommender systems[C]//Proceedings of the 21th ACMSIGKDD International Conference on Knowledge Discovery and Data Mining.ACM,2015:1235-1244.
[5]Paradarami T K,Bastian N D,Wightman J L.A hybrid recommender system using artificial neural networks[J].Expert Systems with Applications,2017,83(C):300-313.
[6]Capdevila J,Arias M,Arratia A.GeoSRS:A hybrid social recommender system for geolocated data[J].Information Systems,2016,57(1):111-128.
[7]Karatzoglou A,Baltrunas L,Shi Y.Learning to rank for recommender systems[C]//Proceedings of the 7th ACM Conference on Recommender Systems.ACM,2013:493-494.
[8]Shi Y,Larson M,Hanjalic A.List-wise learning to rank with matrix factorization for collaborative filtering[C]//Proceedings of the Fourth ACM Conference on Recommender Systems.ACM,2010:269-272.
[9]Gray-DaviesT,Holmes C C,Caron F.Scalable Bayesian nonparametric regression via a plackett-Luce model for conditional ranks[J].Electronic Journal of Statistics,2016,10(2):1807-1828.
[10]Weimer M,Karatzoglou A,Le Q V,et al.COFI RANK,Maximum margin matrix factorization for collaborative ranking[C]//International Conference on Neural Information Processing Systems.Curran Associates Inc,2008:1593-1600.
[11]Yao W,He J,Huang G,et al.SoRank:Incorporating social information into learning to rank models for recommendation[C]//Proceedings of the 23rd International Conference on World Wide Web.ACM,2014:409-410.
[12]Kumar V,Pujari A K,Sahu S K,et al.Proximal maximum margin matrix factorization for collaborative filtering[J].Pattern Recognition Letters,2017,86(1):62-67.