《计算机应用》唯一官方网站 ›› 2024, Vol. 44 ›› Issue (3): 760-766.DOI: 10.11772/j.issn.1001-9081.2023030362

• 数据科学与技术 • 上一篇    下一篇

基于新奇度量的社交事件推荐方法

孙滔1,2, 段张甜3, 朱浩楠4, 郭沛豪5, 孙鹤立5()   

  1. 1.西安交通大学 计算机科学与技术学院, 西安 710049
    2.陕西省科学技术情报研究院, 西安 710054
    3.中国航空工业集团有限公司 西安航空计算技术研究所, 西安 710076
    4.西安交通大学 软件学院, 西安 710049
    5.西安交通大学 新闻与新媒体学院, 西安 710049
  • 收稿日期:2023-04-06 修回日期:2023-05-29 接受日期:2023-05-31 发布日期:2023-10-23 出版日期:2024-03-10
  • 通讯作者: 孙鹤立
  • 作者简介:孙滔(1977—),男,陕西西安人,高级工程师,硕士研究生,主要研究方向:推荐系统、数据分析
    段张甜(1997—),女,陕西西安人,硕士,主要研究方向:数据挖掘、推荐系统
    朱浩楠(1998—),男,陕西榆林人,硕士研究生,主要研究方向:推荐系统、机器学习
    郭沛豪(1999—),男,山东威海人,硕士研究生,主要研究方向:社交网络、新媒体;
  • 基金资助:
    国家自然科学基金资助项目(62072365);传播内容认知国家重点实验室项目(A02105)

Social event recommendation method based on unexpectedness metric

Tao SUN1,2, Zhangtian DUAN3, Haonan ZHU4, Peihao GUO5, Heli SUN5()   

  1. 1.School of Computer Science and Technology,Xi’an Jiaotong University,Xi’an Shaanxi 710049,China
    2.Shaanxi Institute of Science and Technology Information,Xi’an Shaanxi 710054,China
    3.Xi’an Aeronautics Computing Technique Research Institute,Aviation Industry Corporation of China,Limited,Xi’an Shaanxi 710076,China
    4.School of Software Engineering,Xi’an Jiaotong University,Xi’an Shaanxi 710049,China
    5.School of Journalism and New Media,Xi’an Jiaotong University,Xi’an Shaanxi 710049,China
  • Received:2023-04-06 Revised:2023-05-29 Accepted:2023-05-31 Online:2023-10-23 Published:2024-03-10
  • Contact: Heli SUN
  • About author:SUN Tao, born in 1977, M. S. candidate, senior engineer. His research interests include recommendation system, data analysis.
    DUAN Zhangtian, born in 1997, M. S. Her research interests include data mining, recommendation system.
    ZHU Haonan, born in 1998, M. S. candidate. His research interests include recommendation system, machine learning.
    GUO Peihao, born in 1999, M. S. candidate. His research interests include social network, new media.
  • Supported by:
    National Natural Science Foundation of China(62072365);Foundation of State Key Laboratory of Communication Content Cognition(A02105)

摘要:

在社交事件网络(EBSN)中,推荐工作都是从用户的历史喜好出发建模用户偏好,阻碍了用户接触新事物的范围和途径。针对上述问题,提出基于新奇度量的社交事件推荐模型UER(Unexpectedness-based Event Recommendation)。UER模型包括Base和Unexpected两个子模型,首先,Base子模型基于用户、事件以及用户历史事件交互序列特征,通过注意力机制衡量事件在用户历史喜好中的权重,最终预测用户参加事件的概率;其次,Unexpected子模型通过自注意力机制提取用户的多个兴趣表示来计算用户自身新奇度和候选事件对用户的新奇值,从而衡量推荐事件的新奇程度。在Meetup-加州数据集上,UER模型相较于DIN(Deep Interest Network)和PURS(Personalized Unexpected Recommender System)的推荐命中率(HR)分别提高22.9%和30.3%,归一化折损累积收益(NDCG)分别提高27.5%和42.3%,推荐事件的新奇程度分别提高54.5%和21.4%;在Meetup-纽约数据集上,UER模型相较于DIN和PURS的HR分别提高18.2%和21.8%,NDCG分别提高26.9%和32.0%,推荐事件的新奇程度分别提高52.6%和20.8%。

关键词: 社交事件网络, 事件推荐, 异构信息网络, 注意力机制, 交互序列

Abstract:

In Event-Based Social Network (EBSN), the recommendation work starts from the user historical preferences to model user preferences, which hinders the scope and ways for users to access new things. Aiming at the above problems, an unexpectedness metric-based social event recommendation model was proposed, namely UER(Unexpectedness-based Event Recommendation). UER model included two sub-models, Base and Unexpected. Firstly, based on the interaction sequence characteristics of users, events, and user historical events, the Base sub-model used the attention mechanism to measure the weights of events in user historical preferences, and finally predicted the probabilities of users participating in events. Secondly, multiple interest representations of the user were extracted by Unexpected sub-model through the self-attention mechanism to calculate the unexpectedness of the user itself and the unexpectedness value of the candidate event to the user according to the multiple interest representations of the user, so as to measure the unexpectedness of the recommended event. Experimental results on Meetup-California dataset show that compared with Deep Interest Network (DIN) and Personalized Unexpected Recommender System (PURS), the recommendation Hit Ratio (HR) of the UER model is increased by 22.9% and 30.3%, the Normalized Discounted Cumulative Gain (NDCG) is increased by 27.5% and 42.3%, and the unexpectedness of recommended events is increased by 54.5% and 21.4% respectively. On Meetup-NewYork dataset, the recommendation HR of the UER model is increased by 18.2% and 21.8%, the NDCG is increased by 26.9% and 32.0%, and the unexpectedness of recommended events is increased by 52.6% and 20.8% respectively.

Key words: Event-Based Social Network (EBSN), event recommendation? Heterogeneous Information Network (HIN), attention mechanism, interaction sequence

中图分类号: