Journal of Computer Applications ›› 2024, Vol. 44 ›› Issue (10): 3090-3096.DOI: 10.11772/j.issn.1001-9081.2023091352

• Data science and technology • Previous Articles     Next Articles

Sequential recommendation based on hierarchical filter and temporal convolution enhanced self-attention network

Xingyao YANG1(), Hongtao SHEN1, Zulian ZHANG2, Jiong YU1, Jiaying CHEN1, Dongxiao WANG1   

  1. 1.School of Software,Xinjiang University,Urumqi Xinjiang 830091,China
    2.Xinjiang Xinnong Network Information Center,Meteorological Bureau of Xinjiang Uygur Autonomous Region,Urumqi Xinjiang 830002,China
  • Received:2023-10-07 Revised:2023-12-04 Accepted:2023-12-05 Online:2023-12-20 Published:2024-10-10
  • Contact: Xingyao YANG
  • About author:SHEN Hongtao, born in 2001, M. S. candidate. His research interests include recommender system.
    ZHANG Zulian, born in 1984, M. S., senior engineer. Her research interests include numerical prediction, information retrieval.
    YU Jiong, born in 1964, Ph. D., professor. His research interests include grid computing, parallel computing.
    CHEN Jiaying, born in 1988, Ph. D., associate professor. Her research interests include recommender system, social network, data mining.
    WANG Dongxiao, born in 1970, experimentalist. His research interests include cyber security.
  • Supported by:
    Natural Science Foundation of Xinjiang Uygur Autonomous Region(2023D01C17);National Natural Science Foundation of China(62262064);Xinjiang Meteorological Bureau Guidance Project(YD202212);Xinjiang Uygur Autonomous Region Science and Technology Program(2023D4012)

基于层级过滤器和时间卷积增强自注意力网络的序列推荐

杨兴耀1(), 沈洪涛1, 张祖莲2, 于炯1, 陈嘉颖1, 王东晓1   

  1. 1.新疆大学 软件学院,乌鲁木齐 830091
    2.新疆维吾尔自治区气象局 新疆兴农网信息中心,乌鲁木齐 830002
  • 通讯作者: 杨兴耀
  • 作者简介:杨兴耀(1984—),男,湖北襄阳人,副教授,博士,CCF会员,主要研究方向:推荐系统、大数据、信任计算 yangxy@xju.edu.cn
    沈洪涛(2001—),男,安徽淮南人,硕士研究生,主要研究方向:推荐系统
    张祖莲(1984—),女,湖北襄阳人,高级工程师,硕士,主要研究方向:数值预报、信息检索
    于炯(1964—),男,新疆乌鲁木齐人,教授,博士,CCF会员,主要研究方向:网格计算、并行计算
    陈嘉颖(1988—),女,新疆沙湾人,副教授,博士,CCF会员,主要研究方向:推荐系统、社交网络、数据挖掘
    王东晓(1970—),男,黑龙江绥化人,实验师,主要研究方向:网络安全。
  • 基金资助:
    新疆维吾尔自治区自然科学基金资助项目(2023D01C17);国家自然科学基金资助项目(62262064);新疆气象局引导项目(YD202212);新疆维吾尔自治区科技计划项目(2023D4012)

Abstract:

Aiming at the problem of noise arising from user’s unexpected interactions in practical recommendation scenarios and the challenge of capturing short-term demand biases due to the dispersed attention in self-attention mechanism, a model namely FTARec (sequential Recommendation based on hierarchical Filter and Temporal convolution enhanced self-Attention network) was proposed. Firstly, hierarchical filter was used to filter noise in the original data. Then, user embeddings were obtained by combining temporal convolution enhanced self-attention networks with decoupled hybrid location encoding. The deficiencies in modeling short-term dependencies among items were supplemented by enhancing the self-attention network with temporal convolution in this process. Finally, contrastive learning was incorporated to refine user embeddings and predictions were made based on the final user embeddings. Compared to existing sequential recommendation models such as the Self-Attentive Sequential Recommendation (SASRec) and the Filter-enhanced Multi-Layer Perceptron approach for sequential Recommendation (FMLP-Rec), FTARec achieves higher Hit Rate (HR) and Normalized Discounted Cumulative Gain (NDCG) on three publicly available datasets: Beauty, Clothing, and Sports. Compared with the suboptimal DuoRec, FTARec has the HR@10 increased by 7.91%, 13.27%, 12.84%, and the NDCG@10 increased by 5.52%, 8.33%, 9.88%, respectively, verifying the effectiveness of the proposed model.

Key words: self-attention mechanism, filtering algorithm, temporal convolutional network, sequential recommendation, contrastive learning

摘要:

针对实际推荐场景中用户意外交互产生的噪声问题,以及自注意力机制中注意力分布分散导致用户短期需求偏移难以捕获的问题,提出一种基于层级过滤器和时间卷积增强自注意力网络的序列推荐(FTARec)模型。首先,通过层级过滤器过滤原始数据中的噪声;其次,结合时间卷积增强自注意力网络和解耦混合位置编码获取用户嵌入,该过程通过时间卷积增强补充自注意力网络在项目短期依赖建模上的不足;最后,结合对比学习改善用户嵌入,并根据最终用户嵌入进行预测。相较于自注意力序列推荐(SASRec)、过滤增强的多层感知器序列推荐方法(FMLP-Rec)等现有序列推荐模型,FTARec在3个公开数据集Beauty、Clothing和Sports上取得了更高的命中率(HR)和归一化折损累计增益(NDCG),相较于次优的DuoRec,HR@10分别提高了7.91%、13.27%和12.84%,NDCG@10分别提高了5.52%、8.33%和9.88%,验证了所提模型的有效性。

关键词: 自注意力机制, 过滤算法, 时间卷积网络, 序列推荐, 对比学习

CLC Number: