Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations
Graph convolutional networks are widely used in recommendation tasks owing to their ability to learn user and item embeddings using collaborative signals from high-order neighborhoods. Most of the graph convolutional recommendation tasks in existing studies have specialized in modeling a single type...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-08-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/10/16/2956 |
| _version_ | 1827599185186127872 |
|---|---|
| author | Wenhao Zhu Yujun Xie Qun Huang Zehua Zheng Xiaozhao Fang Yonghui Huang Weijun Sun |
| author_facet | Wenhao Zhu Yujun Xie Qun Huang Zehua Zheng Xiaozhao Fang Yonghui Huang Weijun Sun |
| author_sort | Wenhao Zhu |
| collection | DOAJ |
| description | Graph convolutional networks are widely used in recommendation tasks owing to their ability to learn user and item embeddings using collaborative signals from high-order neighborhoods. Most of the graph convolutional recommendation tasks in existing studies have specialized in modeling a single type of user–item interaction preference. Meanwhile, graph-convolution-network-based recommendation models are prone to over-smoothing problems when stacking increased numbers of layers. Therefore, in this study we propose a multi-behavior recommendation method based on graph transformer collaborative filtering. This method utilizes an unsupervised subgraph generation model that divides users with similar preferences and their interaction items into subgraphs. Furthermore, it fuses multi-headed attention layers with temporal coding strategies based on the user–item interaction graphs in the subgraphs such that the learned embeddings can reflect multiple user–item relationships and the potential for dynamic interactions. Finally, multi-behavior recommendation is performed by uniting multi-layer embedding representations. The experimental results on two real-world datasets show that the proposed method performs better than previously developed systems. |
| first_indexed | 2024-03-09T04:08:04Z |
| format | Article |
| id | doaj.art-c6b194750aa640bfb6ce8ace5079a2e3 |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-09T04:08:04Z |
| publishDate | 2022-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-c6b194750aa640bfb6ce8ace5079a2e32023-12-03T14:03:36ZengMDPI AGMathematics2227-73902022-08-011016295610.3390/math10162956Graph Transformer Collaborative Filtering Method for Multi-Behavior RecommendationsWenhao Zhu0Yujun Xie1Qun Huang2Zehua Zheng3Xiaozhao Fang4Yonghui Huang5Weijun Sun6School of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaSchool of Automation, Guangdong University of Technology, Guangzhou 510006, ChinaGraph convolutional networks are widely used in recommendation tasks owing to their ability to learn user and item embeddings using collaborative signals from high-order neighborhoods. Most of the graph convolutional recommendation tasks in existing studies have specialized in modeling a single type of user–item interaction preference. Meanwhile, graph-convolution-network-based recommendation models are prone to over-smoothing problems when stacking increased numbers of layers. Therefore, in this study we propose a multi-behavior recommendation method based on graph transformer collaborative filtering. This method utilizes an unsupervised subgraph generation model that divides users with similar preferences and their interaction items into subgraphs. Furthermore, it fuses multi-headed attention layers with temporal coding strategies based on the user–item interaction graphs in the subgraphs such that the learned embeddings can reflect multiple user–item relationships and the potential for dynamic interactions. Finally, multi-behavior recommendation is performed by uniting multi-layer embedding representations. The experimental results on two real-world datasets show that the proposed method performs better than previously developed systems.https://www.mdpi.com/2227-7390/10/16/2956recommendation systemgraph convolutional networksubgraphtransformermulti-behavior recommendation |
| spellingShingle | Wenhao Zhu Yujun Xie Qun Huang Zehua Zheng Xiaozhao Fang Yonghui Huang Weijun Sun Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations Mathematics recommendation system graph convolutional network subgraph transformer multi-behavior recommendation |
| title | Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations |
| title_full | Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations |
| title_fullStr | Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations |
| title_full_unstemmed | Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations |
| title_short | Graph Transformer Collaborative Filtering Method for Multi-Behavior Recommendations |
| title_sort | graph transformer collaborative filtering method for multi behavior recommendations |
| topic | recommendation system graph convolutional network subgraph transformer multi-behavior recommendation |
| url | https://www.mdpi.com/2227-7390/10/16/2956 |
| work_keys_str_mv | AT wenhaozhu graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT yujunxie graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT qunhuang graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT zehuazheng graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT xiaozhaofang graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT yonghuihuang graphtransformercollaborativefilteringmethodformultibehaviorrecommendations AT weijunsun graphtransformercollaborativefilteringmethodformultibehaviorrecommendations |