Dual Preference Distribution Learning for Item Recommendation

被引:1
作者
Dong, Xue [1 ]
Song, Xuemeng [2 ]
Zheng, Na [3 ]
Wei, Yinwei [4 ]
Zhao, Zhongzhou [5 ]
机构
[1] Shandong Univ, Sch Software, 1500 Shunhua Rd, Jinan 250101, Peoples R China
[2] Shandong Univ, Sch Comp Sci & Technol, 72 Binhai Rd, Qingdao 266237, Peoples R China
[3] Natl Univ Singapore, Inst Data Sci, 21 Lower Kent Ridge Rd, Singapore, Singapore
[4] Natl Univ Singapore, Sch Comp, 21 Lower Kent Ridge Rd, Singapore, Singapore
[5] DAMO Acad, Alibaba Grp, 969 Yuhang Westen Rd, Hangzhou 311121, Zhejiang, Peoples R China
基金
国家重点研发计划;
关键词
Recommender system; preference distribution learning; explainable recommendation;
D O I
10.1145/3565798
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Recommender systems can automatically recommend users with items that they probably like. The goal of them is to model the user-item interaction by effectively representing the users and items. Existing methods have primarily learned the user's preferences and item's features with vectorized embeddings, and modeled the user's general preferences to items by the interaction of them. In fact, users have their specific preferences to item attributes and different preferences are usually related. Therefore, exploring the fine-grained preferences as well as modeling the relationships among user's different preferences could improve the recommendation performance. Toward this end, we propose a dual preference distribution learning framework (DUPLE), which aims to jointly learn a general preference distribution and a specific preference distribution for a given user, where the former corresponds to the user's general preference to items and the latter refers to the user's specific preference to item attributes. Notably, the mean vector of each Gaussian distribution can capture the user's preferences, and the covariance matrix can learn their relationship. Moreover, we can summarize a preferred attribute profile for each user, depicting his/her preferred item attributes. We then can provide the explanation for each recommended item by checking the overlap between its attributes and the user's preferred attribute profile. Extensive quantitative and qualitative experiments on six public datasets demonstrate the effectiveness and explainability of the DUPLE method.
引用
收藏
页数:22
相关论文
共 63 条
[1]   Learning Heterogeneous Knowledge Base Embeddings for Explainable Recommendation [J].
Ai, Qingyao ;
Azizi, Vahid ;
Chen, Xu ;
Zhang, Yongfeng .
ALGORITHMS, 2018, 11 (09)
[2]  
[Anonymous], 2021, INT ACM SIGIR C RES, DOI DOI 10.1145/3404835.3462893
[3]  
[Anonymous], 2018, SIGIR C RES DEV INF, DOI DOI 10.1145/3209978.3210083
[4]  
[Anonymous], 2020, ACM INT C INF KNOWL, DOI DOI 10.1145/3340531.3412038
[5]  
[Anonymous], 2019, ACM SIGKDD INT C KNO, DOI DOI 10.1145/3292500.3330989
[6]   Neural Attentional Rating Regression with Review-level Explanations [J].
Chen, Chong ;
Zhang, Min ;
Liu, Yiqun ;
Ma, Shaoping .
WEB CONFERENCE 2018: PROCEEDINGS OF THE WORLD WIDE WEB CONFERENCE (WWW2018), 2018, :1583-1592
[7]  
Chen Yifan, 2020, ACM T INF SYST, V38
[8]  
Chen ZX, 2019, PROCEEDINGS OF THE TWENTY-EIGHTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, P2137
[9]   MMALFM: Explainable Recommendation by Leveraging Reviews and Images [J].
Cheng, Zhiyong ;
Chang, Xiaojun ;
Zhu, Lei ;
Kanjirathinkal, Rose C. ;
Kankanhalli, Mohan .
ACM TRANSACTIONS ON INFORMATION SYSTEMS, 2019, 37 (02)
[10]  
Costa F.J, 2018, PROC SBMICRO 2018, P1