Label-informed Graph Structure Learning for Node Classification

TL;DR

This paper introduces a label-informed graph structure learning framework that explicitly incorporates label information via a class transition matrix, improving node classification performance.

Contribution

It proposes a novel method that leverages label information in graph structure learning, addressing the limitations of previous feature-only approaches.

Findings

Outperforms or matches state-of-the-art baselines on seven datasets.

Effectively incorporates label information into graph structure learning.

Enhances node classification accuracy.

Abstract

Graph Neural Networks (GNNs) have achieved great success among various domains. Nevertheless, most GNN methods are sensitive to the quality of graph structures. To tackle this problem, some studies exploit different graph structure learning strategies to refine the original graph structure. However, these methods only consider feature information while ignoring available label information. In this paper, we propose a novel label-informed graph structure learning framework which incorporates label information explicitly through a class transition matrix. We conduct extensive experiments on seven node classification benchmark datasets and the results show that our method outperforms or matches the state-of-the-art baselines.