Dual-level graph contrastive collaborative filtering

Abstract The latest research positions graph-based collaborative filtering as an effective strategy in recommendation systems, enabling the analysis of user preferences via user-item interaction graphs. However, such methods often struggle with data sparsity issues in real-world scenarios. To addres...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiahao Wang, Qingshuai Wang, Kai Ma, Noor Farizah Ibrahim, Zurinahni Zainol
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-10621-x
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849234963956760576
author Jiahao Wang
Qingshuai Wang
Kai Ma
Noor Farizah Ibrahim
Zurinahni Zainol
author_facet Jiahao Wang
Qingshuai Wang
Kai Ma
Noor Farizah Ibrahim
Zurinahni Zainol
author_sort Jiahao Wang
collection DOAJ
description Abstract The latest research positions graph-based collaborative filtering as an effective strategy in recommendation systems, enabling the analysis of user preferences via user-item interaction graphs. However, such methods often struggle with data sparsity issues in real-world scenarios. To address this, contrastive learning mechanisms have been integrated into graph collaborative filtering, though existing approaches are limited to single-view designs at either the graph or node level, restricting overall model performance. In response, we propose an innovative dual-level graph contrastive collaborative filtering method (DL-GCL) that combines both graph and node-level views. First, at the node level, we employ a matrix decomposition technique during the preprocessing phase to decompose and reconstruct the bipartite graph. Based on the reconstructed results, contrastive views are constructed to capture local collaborative information. Subsequently, considering the potential noise introduced by node-level views, we mitigate the impact of uncertain noise by capturing the model’s maximum gradient state at the graph level. Using the Fast Gradient Sign Method (FGSM), we perturb the model’s representation vectors under worst-case conditions, thereby mitigating noise from node-level views and extracting global collaborative information. Finally, DL-GCL employs a multi-task learning strategy to optimize local-global views and BPR (Bayesian Personalized Ranking) loss functions. Through extensive experiments on four public datasets, the evaluation metrics NDCG and Recall show up to a 24.5% improvement compared to the latest graph contrastive models. This highlights the strong performance of DL-GCL in improving recommendation system robustness and mitigating data sparsity.
format Article
id doaj-art-fd6e5c8d0a7e4c7ba6ec905f754c71ae
institution Kabale University
issn 2045-2322
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-fd6e5c8d0a7e4c7ba6ec905f754c71ae2025-08-20T04:02:57ZengNature PortfolioScientific Reports2045-23222025-07-0115111910.1038/s41598-025-10621-xDual-level graph contrastive collaborative filteringJiahao Wang0Qingshuai Wang1Kai Ma2Noor Farizah Ibrahim3Zurinahni Zainol4School of Computer Sciences, Universiti Sains MalaysiaSchool of Computer Sciences, Universiti Sains MalaysiaSchool of Computer Sciences, Universiti Sains MalaysiaSchool of Computer Sciences, Universiti Sains MalaysiaSchool of Computer Sciences, Universiti Sains MalaysiaAbstract The latest research positions graph-based collaborative filtering as an effective strategy in recommendation systems, enabling the analysis of user preferences via user-item interaction graphs. However, such methods often struggle with data sparsity issues in real-world scenarios. To address this, contrastive learning mechanisms have been integrated into graph collaborative filtering, though existing approaches are limited to single-view designs at either the graph or node level, restricting overall model performance. In response, we propose an innovative dual-level graph contrastive collaborative filtering method (DL-GCL) that combines both graph and node-level views. First, at the node level, we employ a matrix decomposition technique during the preprocessing phase to decompose and reconstruct the bipartite graph. Based on the reconstructed results, contrastive views are constructed to capture local collaborative information. Subsequently, considering the potential noise introduced by node-level views, we mitigate the impact of uncertain noise by capturing the model’s maximum gradient state at the graph level. Using the Fast Gradient Sign Method (FGSM), we perturb the model’s representation vectors under worst-case conditions, thereby mitigating noise from node-level views and extracting global collaborative information. Finally, DL-GCL employs a multi-task learning strategy to optimize local-global views and BPR (Bayesian Personalized Ranking) loss functions. Through extensive experiments on four public datasets, the evaluation metrics NDCG and Recall show up to a 24.5% improvement compared to the latest graph contrastive models. This highlights the strong performance of DL-GCL in improving recommendation system robustness and mitigating data sparsity.https://doi.org/10.1038/s41598-025-10621-xGraph neural networksContrastive learningRecommender systemsCollaborative filtering
spellingShingle Jiahao Wang
Qingshuai Wang
Kai Ma
Noor Farizah Ibrahim
Zurinahni Zainol
Dual-level graph contrastive collaborative filtering
Scientific Reports
Graph neural networks
Contrastive learning
Recommender systems
Collaborative filtering
title Dual-level graph contrastive collaborative filtering
title_full Dual-level graph contrastive collaborative filtering
title_fullStr Dual-level graph contrastive collaborative filtering
title_full_unstemmed Dual-level graph contrastive collaborative filtering
title_short Dual-level graph contrastive collaborative filtering
title_sort dual level graph contrastive collaborative filtering
topic Graph neural networks
Contrastive learning
Recommender systems
Collaborative filtering
url https://doi.org/10.1038/s41598-025-10621-x
work_keys_str_mv AT jiahaowang duallevelgraphcontrastivecollaborativefiltering
AT qingshuaiwang duallevelgraphcontrastivecollaborativefiltering
AT kaima duallevelgraphcontrastivecollaborativefiltering
AT noorfarizahibrahim duallevelgraphcontrastivecollaborativefiltering
AT zurinahnizainol duallevelgraphcontrastivecollaborativefiltering