Abstract | ||
---|---|---|
Online social platforms have provided a large amount of available information to recommendation systems. With this intuition, social recommendation systems emerged and have attracted increasing attention over the past years. Most existing social recommendation methods only use explicit social relationships among users. However, implicit social relationships can effectively improve the quality of recommendation when users only have few social relationships. To this end, the discovery of implicit relations among users plays a central role in advancing social recommendation. In this paper, we propose a novel approach to fuse direct and indirect friends toward discovering more accurate social recommendation method. We learn users' preferences by carefully integrating users' direct and indirect friends. In particular, we construct item rankings based on the feedback from users' direct and indirect friends on the item. Furthermore, to distinguish the impact of users' direct friends and indirect friends, we also extend the ranking assumption in item domain to user domain, so that information from user rankings can be leveraged to further improve the recommendation performance. Extensive experiments on two real-world datasets demonstrate the effectiveness of the proposed method. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1007/978-3-030-75765-6_31 | ADVANCES IN KNOWLEDGE DISCOVERY AND DATA MINING, PAKDD 2021, PT II |
Keywords | DocType | Volume |
Social information, Indirect friends, Item ranking, User ranking | Conference | 12713 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yunhe Wei | 1 | 0 | 0.34 |
Huifang Ma | 2 | 290 | 29.69 |
Ruoyi Zhang | 3 | 0 | 0.34 |
Zhixin Li | 4 | 12 | 19.62 |
Liang Chang | 5 | 23 | 14.22 |