Rec-GNN: Research on Social Recommendation based on Graph Neural Networks

DOI

10.2991/978-94-6463-108-1_55How to use a DOI?

Keywords

Graph Neural network; Recommendation system; Social recommendation; Neural network

Abstract

Solve the problem that the accuracy of scoring prediction is not high due to insufficient learning of the features of two graph data (user social graph and user item graph) in the social recommendation system (GraphRec), and improve the accuracy of the social recommendation system. In this paper, the user associated with item and the user evaluation of items are improved, and the commodity category attribute is added; And use users as a bridge to connect user-user social graph and user-item graph; Analyze the heterogeneous advantages between nodes, and make full use of the two graph data contained in the social recommendation system to generate the feature vectors that are sufficiently differentiated and representative for users and goods. Compared with the GraphRec basic model without commodity category attribute, the experimental results show that the improved (Rec-GNN) based social recommendation model has a higher accuracy rate in scoring prediction. In this method, RMSE indicators have been improved by 2.5% and 1.5%, respectively, on Ciao and Epinions' two real data sets, and MAE indicators have been improved by 1.3% and 1.3%, respectively. The experimental performance of the model after adding a deeper potential relationship between users and commodities was not tested. This paper analyzes the complex interaction between nodes and adds rich feature infor-mation, such as user social information, user ratings on items, and item category attributes, to obtain a more differentiated feature vector representation of users and goods, which has higher accuracy in the social recommendation system score prediction.

Copyright

© 2022 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Gaofei Si
AU  - Shuwei Xu
AU  - Zhaoke Li
AU  - Jingyun Zhang
PY  - 2022
DA  - 2022/12/30
TI  - Rec-GNN: Research on Social Recommendation based on Graph Neural Networks
BT  - Proceedings of the 2022 International Conference on Computer Science, Information Engineering and Digital Economy (CSIEDE 2022)
PB  - Atlantis Press
SP  - 478
EP  - 485
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-108-1_55
DO  - 10.2991/978-94-6463-108-1_55
ID  - Si2022
ER  -