Abstract:Fusion of multi-source information is one of the primary methods to alleviate data sparsity in recommender systems (RS). Hypergraphs have shown remarkable capabilities in dealing with the diversity of multi-source information, especially in modeling high-order user-item relationships. However, most hypergraph research focuses on constructing hypergraphs using a particular type of hyperedge, which might not fully capture the high-order implicit associations among heterogeneous information sources. Furthermore, existing Hypergraph Attention Networks (Hyper-GAT), mainly emphasizes on information propagation between nodes and hyperedges, insufficiently exploring density information within complex hyperedges. Moreover, when hypergraphs combine data from multiple heterogeneous graphs, redundant information from different viewpoints can reduce the effectiveness of multivariate modeling hyperedges. Thus, we propose a recommendation algorithm based on multi-source heterogeneous hypergraphs and contrastive learning (MHCLR), which improves recommendation accuracy through multi-source information fusion and higher-order information correlation. First, multi-source heterogeneous hypergraphs (MHC) are generated by combining distance, behavior, attribute, and prediction hyperedges. We mine associations among multi-source information, enhancing high-order semantic connections among users and items. Then, a Spatial Density Hypergraph Attention Network (SD-HGAT) is proposed based on composite hyperedges, which enriches the user and item embedding representations by focusing on nodes and hyperedges density. Finally, we design a multiple cross view contrastive learning (MCC) that compares views centered around knowledge graphs with hypergraphs, improving the accuracy of multivariate relationship modeling and enabling multi-level user profiling construction. It is observed that MHCLR outperforms the baselines in terms of Recall, Precision, and NDCG based on the experimental results by Yelp, Last-FM, and Douban.

HAMF: Highly Adaptable Multi-graph Fusion for Cross-domain Graphs

Hamming Embedding Sensitivity Guided Fusion Network for 3D Shape Representation.

Multi-Graph Heterogeneous Interaction Fusion for Social Recommendation

Attributed Graph Clustering Network with Adaptive Feature Fusion

GraphTransfer: A Generic Feature Fusion Framework for Collaborative Filtering

Meta-Graph Based Recommendation Fusion over Heterogeneous Information Networks.

HAMUR: Hyper Adapter for Multi-Domain Recommendation

Fusing heterogeneous information for multi-modal attributed network embedding

Side Information Fusion for Recommender Systems over Heterogeneous Information Network

A Recommendation Method Based on Multi-Source Heterogeneous Hypergraphs and Contrastive Learning

A Relation-Aware Heterogeneous Graph Transformer on Dynamic Fusion for Multimodal Classification Tasks

Anchor Graph Structure Fusion Hashing for Cross-Modal Similarity Search

MHGNN: Multi-view fusion based Heterogeneous Graph Neural Network

MHNF: Multi-hop Heterogeneous Neighborhood information Fusion graph representation learning

AF-GCN: Attribute-Fusing Graph Convolution Network for Recommendation

HGAMLP: Heterogeneous Graph Attention MLP with De-Redundancy Mechanism

Data fusing driven graph embedding model for multi-source heterogeneous knowledge graph

DM-Fusion: Deep Model-Driven Network for Heterogeneous Image Fusion.

A Heterogeneous Graph Based Framework for Multimodal Neuroimaging Fusion Learning

Multiple heterogeneous network representation learning based on multi-granularity fusion

An interlayer feature fusion-based heterogeneous graph neural network