融合分类与协同过滤的情境感知音乐推荐算法
|
摘要
针对用户情境信息,提出一种融合分类与协同过滤的情境感知音乐推荐算法.首先,通过计算用户情境信息的相似度,由协同过滤算法得到初始音乐推荐列表;然后通过机器学习算法训练分类模型,得出用户在特定情境下的音乐类型偏好;最后将协同过滤得到的推荐列表与分类模型得到的音乐类型偏好进行融合,为特定情境的用户提供个性化音乐推荐.该算法不仅有效地降低了推荐过程的复杂度,还使传统的协同过滤推荐算法具备了情境感知的能力.实验结果表明,该方法可以有效地提高个性化音乐推荐系统的性能.
A context-aware music recommendation algorithm combining classification and collaborative filtering was proposed to solve the problem of how to use context information for personalized recommendation. First, this paper was adopted a method of calculating the similarity of user context information, and was used the collaborative filtering algorithm to obtain the initial music recommendation list. Then the classification model had been trained by the machine learning algorithm to derive the user's music type preference in a specific context. The recommended list generated by collaborative filtering algorithm was merged with the music type preference predicted by the classifier to provide personalized music recommendations for users in specific situations. The algorithm not only effectively reduced the complexity of the recommendation process, but also enabled the traditional collaborative filtering recommendation algorithm to have context-awareness. Experimental results had proved that this method can effectively improve the performance of personalized music recommendation system.
A context-aware music recommendation algorithm combining classification and collaborative filtering was proposed to solve the problem of how to use context information for personalized recommendation. First, this paper was adopted a method of calculating the similarity of user context information, and was used the collaborative filtering algorithm to obtain the initial music recommendation list. Then the classification model had been trained by the machine learning algorithm to derive the user's music type preference in a specific context. The recommended list generated by collaborative filtering algorithm was merged with the music type preference predicted by the classifier to provide personalized music recommendations for users in specific situations. The algorithm not only effectively reduced the complexity of the recommendation process, but also enabled the traditional collaborative filtering recommendation algorithm to have context-awareness. Experimental results had proved that this method can effectively improve the performance of personalized music recommendation system.
引文
[1] 谭学清,何珊.音乐个性化推荐系统研究综述[J].数据分析与知识发现,2014,30(9):22-32.
[2] 熊敏.基于情景感知的个性化音乐推荐算法的研究[D].杭州:浙江理工大学,2016.
[3] WANG X,ROSENBLUM D,WANG Y.Context-aware mobile music recommendation for daily activities[C]//Acm Multimedia.Nara:ACM,2012:99-108.
[4] DOLATKIA I,AZIMZADEH F.Music recommendation system based on the continuous combination of contextual information[C]// Second International Conference on Web Research.Tehran:IEEE,2016:108-114.
[5] JIANG C,HE Y.Smart-DJ:context-aware personalization for music recommendation on smartphones[C]//IEEE International Conference on Parallel & Distributed Systems.Wuhan:IEEE,2017:133-140.
[6] 谭学清,何珊.用户情境下基于信息增益和项目的协同过滤推荐技术研究[J].情报杂志,2014(7):165-170.
[7] 王明佳,韩景倜.基于多维度用户相似性度量的协同过滤推荐算法[J].统计与决策,2018(9):66-69.
[8] POLATIDIS N,GEORGIADIS C K.A multi-level collaborative filtering method that improves recommendations[J].Expert Systems with Applications,2016,48:100-110.
[9] 乔磊.基于多维情境的移动信息服务个性化推荐算法研究[D].北京:北京交通大学,2016.
[10] 罗海媛,章牧.基于用户多属性加权和兴趣相似度的协同过滤算法研究[J].情报探索,2018(5):1-7.
[11] 李想.基于XGBoost算法的多因子量化选股方案策划[D].上海:上海师范大学,2017.
[12] PORTUGAL I,ALENCAR P,COWAN D.The use of machine learning algorithms in recommender systems:a systematic review[J].Expert Systems with Applications,2018,97:205-227.
[13] WANG Y,DENG J,GAO J,et al.A hybrid user similarity model for collaborative filtering[J].Information Sciences,2017,418/419:102-118.
[14] SON L H.Dealing with the new user cold-start problem in recommender systems:a comparative review[J].Information Systems,2016,58:87-104.
[15] 项亮.推荐系统实践[M].北京:人民邮电出版社,2012.
[16] KUNAVER M,PO?RL T.Diversity in recommender systems,a survey[J].Knowledge-Based Systems,2017,123:154-162.
[2] 熊敏.基于情景感知的个性化音乐推荐算法的研究[D].杭州:浙江理工大学,2016.
[3] WANG X,ROSENBLUM D,WANG Y.Context-aware mobile music recommendation for daily activities[C]//Acm Multimedia.Nara:ACM,2012:99-108.
[4] DOLATKIA I,AZIMZADEH F.Music recommendation system based on the continuous combination of contextual information[C]// Second International Conference on Web Research.Tehran:IEEE,2016:108-114.
[5] JIANG C,HE Y.Smart-DJ:context-aware personalization for music recommendation on smartphones[C]//IEEE International Conference on Parallel & Distributed Systems.Wuhan:IEEE,2017:133-140.
[6] 谭学清,何珊.用户情境下基于信息增益和项目的协同过滤推荐技术研究[J].情报杂志,2014(7):165-170.
[7] 王明佳,韩景倜.基于多维度用户相似性度量的协同过滤推荐算法[J].统计与决策,2018(9):66-69.
[8] POLATIDIS N,GEORGIADIS C K.A multi-level collaborative filtering method that improves recommendations[J].Expert Systems with Applications,2016,48:100-110.
[9] 乔磊.基于多维情境的移动信息服务个性化推荐算法研究[D].北京:北京交通大学,2016.
[10] 罗海媛,章牧.基于用户多属性加权和兴趣相似度的协同过滤算法研究[J].情报探索,2018(5):1-7.
[11] 李想.基于XGBoost算法的多因子量化选股方案策划[D].上海:上海师范大学,2017.
[12] PORTUGAL I,ALENCAR P,COWAN D.The use of machine learning algorithms in recommender systems:a systematic review[J].Expert Systems with Applications,2018,97:205-227.
[13] WANG Y,DENG J,GAO J,et al.A hybrid user similarity model for collaborative filtering[J].Information Sciences,2017,418/419:102-118.
[14] SON L H.Dealing with the new user cold-start problem in recommender systems:a comparative review[J].Information Systems,2016,58:87-104.
[15] 项亮.推荐系统实践[M].北京:人民邮电出版社,2012.
[16] KUNAVER M,PO?RL T.Diversity in recommender systems,a survey[J].Knowledge-Based Systems,2017,123:154-162.