Heterogeneous graph neural network with relation-aware label propagation for unbalanced node classification

Chengcheng Sun, Cheng Zhai, Qihan Feng, Xiaobin Rui and Zhixiao Wang

Physica A: Statistical Mechanics and its Applications, 2025, vol. 660, issue C

Abstract: Node classification is one of the core downstream tasks of heterogeneous graph representation learning. However, existing heterogeneous graph neural networks (HGNNs) often exhibit bias toward the majority class, resulting in poor classification performance for the minority classes. Recently, some studies have begun to focus on the imbalance issue in homogeneous graphs. However, due to the inherent heterogeneity and imbalance of heterogeneous graphs, the exploration of imbalanced node classification in heterogeneous graphs remains under-explored. To bridge this gap, this paper investigates the representation learning on heterogeneous graphs and propose a novel model named Heterogeneous Graph Neural Network with Relation-aware Label Propagation (RLP-HGNN). To handle the heterogeneity, we design a relation-aware label propagation to obtain pseudo-labels of nodes in heterogeneous graphs. These pseudo-labels serve as a data augmentation strategy for subsequent phases. Different types of nodes may have different importance, and we adopt dual-level aggregation based on a type-attention mechanism for heterogeneous message passing among different relation subgraphs. To deal with the imbalance issue, we adopt different imbalance strategies to alleviate the classification bias in heterogeneous graphs, including Re-weight, Balanced Softmax, and PC Softmax. By combining relation-aware label propagation and dual-level aggregation into a multi-objective optimization problem, we train the whole model in an end-to-end fashion. We further study the performance of different methods under different imbalance ratio settings. With unbalanced strategies study, ablation analysis, and parameter sensitivity analysis, our experiments on heterogeneous graphs demonstrate the effectiveness and generalizability of our proposed approach in relieving imbalance issues.

Keywords: Heterogeneous graph neural network; Relation-aware label propagation; Unbalanced node classification (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0378437125000214
Full text for ScienceDirect subscribers only. Journal offers the option of making the article available online on Science direct for a fee of $3,000

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:phsmap:v:660:y:2025:i:c:s0378437125000214

DOI: 10.1016/j.physa.2025.130369

Access Statistics for this article

Physica A: Statistical Mechanics and its Applications is currently edited by K. A. Dawson, J. O. Indekeu, H.E. Stanley and C. Tsallis

More articles in Physica A: Statistical Mechanics and its Applications from Elsevier
Bibliographic data for series maintained by Catherine Liu ().