广东工业大学学报 ›› 2023, Vol. 40 ›› Issue (04): 37-44.doi: 10.12052/gdutxb.220009

• 计算机科学与技术 • 上一篇    下一篇

基于用户长短期偏好和音乐情感注意力的音乐推荐模型

吴亚迪, 陈平华   

  1. 广东工业大学 计算机学院, 广东 广州 510006
  • 收稿日期:2022-01-27 出版日期:2023-07-25 发布日期:2023-08-02
  • 通信作者: 陈平华(1967–)，男，教授，主要研究方向为云计算、大数据、推荐系统，E-mail：pinghuachen@163.com
  • 作者简介:吴亚迪 (1997–)， 女，硕士研究生，主要研究方向为推荐系统、数据挖掘，E-mail：603001605@qq.com
  • 基金资助:
    广东省重点领域研发计划项目 (2021B0101200002，2020B0101100001)；广东省科技计划项目 (2020B1010010010)

A Music Recommendation Model Based on Users' Long and Short Term Preferences and Music Emotional Attention

Wu Ya-di, Chen Ping-hua   

  1. School of Computer Science and Technology, Guangdong University of Technology, Guangzhou 510006, China
  • Received:2022-01-27 Online:2023-07-25 Published:2023-08-02

HTML

PDF (PC)

851

摘要： 针对现有音乐推荐在用户偏好建模时忽略用户长期偏好，或对用户记录统一建模时忽略历史信息与当前情境联系的问题，提出一种基于用户长短期偏好和音乐情感注意力的音乐推荐模型。首先将用户听歌记录切分为多个历史序列和当前序列，利用多个长短期记忆网络分别进行特征提取，得到用户长短期偏好：对于历史音乐序列，提出序列时段的概念，并进行序列时段加权计算，得到长期偏好；对于当前序列，利用平均池化提取当前情景音乐特征，得到短期偏好。其次，从音乐声学信号中学习音乐情感特征，应用注意力机制计算音乐情感因子。最后，将音乐情感因子融入用户长短期偏好，得到一个音乐推荐列表。在Last.fm真实数据集上的实验结果表明，模型的NDCG@10达到了0.5435，优于现有方法；消融实验和特征贡献分析进一步验证了模型的有效性。

关键词: 音乐推荐, 用户偏好, 音乐情感, 长短期记忆网络, 注意力机制

Abstract: Users' long-term preferences and the relationship between historical information and current situation are usually ignored in existing user preference modeling and user record unified modeling. To address this, in this paper, we propose a music recommendation model based on users' long-term and short-term preferences and music emotional attention. Firstly, we divide the user's listening record into multiple historical and current sequences, and learn the features by multiple long and short-term memory networks, respectively, to obtain user's long and short-term preferences. For the historical music sequences, we propose the concept of sequence period is proposed to obtain the long-term preferences by calculating the weights of the sequence period. For the current sequence, we use the average pooling to extract the music features of the current scene to obtain the short-term preference. Secondly, we learn the emotional characteristics of music from the acoustic signals, and use the attention mechanism to calculate the emotional factors of music. Finally, we integrate the music emotion factor into the user's long-term and short-term preference to generate a music recommendation list. The experimental results on the last.fm real data set show that the proposed model achieves 0.5435 in term of the NDCG@10, which is better than the existing methods. The ablation experiment and characteristic contribution analysis further demonstrate the effectiveness of the model.

Key words: music recommendation, user preference, sequence analysis, long-short term memory, attentional mechanism

中图分类号: 

  • TP391
[1] SUN K, QIAN T, CHEN T, et al. Where to go next: modeling long-and short-term user preferences for point-of-interest recommendation[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto: AAAI, 2020: 214-221.
[2] BOGDANOV D, HARO M, FUHRMANN F, et al. Semantic audio content-based music recommendation and visualization based on user preference examples[J]. Information Processing & Management, 2013, 49(1): 13-33.
[3] OORD A, DIELEMAN S, SCHRAUWEN B. Deep content-based music recommendation[J]. Advances in Neural Information Processing Systems, 2013, 26: 2643-2651.
[4] 彭嘉恩, 邓秀勤, 刘太亨, 等. 融合社交和标签信息的隐语义模型推荐算法[J]. 广东工业大学学报, 2018, 35(4): 45-50.PENG J E, DENG X Q, LIU T H, et al. A recommendation algorithm of latent factor model fused with the social and tag information[J]. Journal of Guangdong University of Technology, 2018, 35(4): 45-50.
[5] CHENG R, TANG B. A music recommendation system based on acoustic features and user personalities[C]//Trends and Applications in Knowledge Discovery and Data Mining. Aucland: Springer Cham, 2016: 203-213.
[6] CHENG Z, SHEN J, NIE L, et al. Exploring user-specific information in music retrieval[C]//Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: Association for Computing Machinery, 2017: 655-664.
[7] CHEN X. The application of neural network with convolution algorithm in western music recommendation practice[J]. Journal of Ambient Intelligence and Humanized Computing, 2020, 11(3): 1-11.
[8] ORAMAS S, OSTUNI V C, NOIA T D, et al. Sound and music recommendation with knowledge graphs[J]. ACM Transactions on Intelligent Systems and Technology (TIST), 2016, 8(2): 1-21.
[9] LEE J, LEE K, PARK J, et al. Deep content-user embedding model for music recommendation[EB/OL]. arXiv: 1807.06786(2018-07-18)[2022-01-13].https://doi.org/10.48550/arXiv.1807.06786.
[10] SACHDEVA N, GUPTA K, PUDI V. Attentive neural architecture incorporating song features for music recommendation[C]//Proceedings of the 12th ACM Conference on Recommender Systems. New York: Association for Computing Machinery, 2018: 417-421.
[11] 杨明极, 刘畅, 宋泽. 注意力机制与改进RNN的混合音乐推荐算法研究[J]. 小型微型计算机系统, 2020, 41(10): 2235-2240.YANG M J, LIU C, SONG Z. Research on music recommendation algorithm based on attention mechanism and improved RNN[J]. Journal of Chinese Computer Systems, 2020, 41(10): 2235-2240.
[12] WANG D, DENG S, ZHANG X, et al. Learning to embed music and metadata for context-aware music recommendation[J]. World Wide Web, 2018, 21(5): 1399-1423.
[13] SHEN T, JIA J, LI Y, et al. PEIA: Personality and emotion integrated attentive model for music recommendation on social media platforms[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto: AAAI, 2020: 206-213.
[14] DHAHRI C, MATSUMOTO K, HOASHI K. Mood-Aware music recommendation via adaptive song embedding[C]//2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM). Barcelona: IEEE, 2018: 135-138.
[15] ABDUL A, CHEN J, LIAO H Y, et al. An emotion-aware personalized music recommendation system using a convolutional neural networks approach[J]. Applied Sciences, 2018, 8(7): 1103.
[16] BHAUMIK M, ATTAH P U, JAVED F. Emotion integrated music recommendation system using generative adversarial networks[J]. SMU Data Science Review, 2021, 5(3): 4.
[17] 赵永建, 杨振国, 刘文印. 基于双向条目注意网络的推荐系统[J]. 广东工业大学学报, 2020, 37(4): 27-34.ZHAO Y J, YANG Z G, LIU S Y. DIAN: dual-aspect item attention network for item-based recommendation[J]. Journal of Guangdong University of Technology, 2020, 37(4): 27-34.
[18] GUO H, TANG R, YE Y, et al. DeepFM: a factorization-machine based neural network for CTR prediction[EB/OL]. arXiv: 1703.04247(2017-03-13)[2022-01-13].https://doi.org/10.48550/arXiv.1703.04247.
[19] XIAO J, YE H, HE X, et al. Attentional factorization machines: learning the weight of feature interactions via attention networks[EB/OL]. arXiv: 1708.04617(2017-08-15)[2022-01-13].https://doi.org/10.48550/arXiv.1708.04617.
[20] JIN Y, HAN C. A music recommendation algorithm based on clustering and latent factor model[C]//2019 International Conference on Computer Science Communication and Network Security. Sanya: EDP Sciences, 2020: 1-9.
[1] 曹智雄, 吴晓鸰, 骆晓伟, 凌捷. 融合迁移学习与YOLOv5的安全帽佩戴检测算法[J]. 广东工业大学学报, 2023, 40(04): 67-76.
[2] 赖东升, 冯开平, 罗立宏. 基于多特征融合的表情识别算法[J]. 广东工业大学学报, 2023, 40(03): 10-16.
[3] 吴俊贤, 何元烈. 基于通道注意力的自监督深度估计方法[J]. 广东工业大学学报, 2023, 40(02): 22-29.
[4] 刘洪伟, 林伟振, 温展明, 陈燕君, 易闽琦. 基于MABM的消费者情感倾向识别模型——以电影评论为例[J]. 广东工业大学学报, 2022, 39(06): 1-9.
[5] 滕少华, 董谱, 张巍. 融合语义结构的注意力文本摘要模型[J]. 广东工业大学学报, 2021, 38(03): 1-8.
[6] 梁观术, 曹江中, 戴青云, 黄云飞. 一种基于注意力机制的无监督商标检索方法[J]. 广东工业大学学报, 2020, 37(06): 41-49.
[7] 张舒, 莫赞, 柳建华, 杨培琛, 刘洪伟. 基于NWD集成算法的多粒度微博用户兴趣画像构建[J]. 广东工业大学学报, 2020, 37(04): 42-50.
[8] 吴家湖, 熊华, 宗睿, 赵曜, 周贤中. 基于循环神经网络的目标转弯机动类型识别[J]. 广东工业大学学报, 2020, 37(02): 67-73.
[9] 曾碧卿, 韩旭丽, 王盛玉, 徐如阳, 周武. 基于双注意力卷积神经网络模型的情感分析研究[J]. 广东工业大学学报, 2019, 36(04): 10-17.
[10] 高俊艳, 刘文印, 杨振国. 结合注意力与特征融合的目标跟踪[J]. 广东工业大学学报, 2019, 36(04): 18-23.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 2017 《广东工业大学学报》编辑部
地址：广州市东风东路729号广东工业大学 邮编：510090
电话：020-37626653,020-37626139,020-37626396
邮箱：xbzrb@gdut.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发