Region-enhanced Deep Graph Convolutional Networks for Rumor Detection

TL;DR

This paper introduces RDGCN, a novel deep graph convolutional network that captures regionalized rumor propagation patterns and enhances rumor detection accuracy on social media data.

Contribution

The study proposes a region-enhanced deep GCN with unsupervised learning of propagation patterns and a source-enhanced residual layer, addressing propagation structure limitations.

Findings

Outperforms baseline models on Twitter15 and Twitter16 datasets.

Improves early rumor detection accuracy.

Effectively captures regionalized propagation patterns.

Abstract

Social media has been rapidly developing in the public sphere due to its ease of spreading new information, which leads to the circulation of rumors. However, detecting rumors from such a massive amount of information is becoming an increasingly arduous challenge. Previous work generally obtained valuable features from propagation information. It should be noted that most methods only target the propagation structure while ignoring the rumor transmission pattern. This limited focus severely restricts the collection of spread data. To solve this problem, the authors of the present study are motivated to explore the regionalized propagation patterns of rumors. Specifically, a novel region-enhanced deep graph convolutional network (RDGCN) that enhances the propagation features of rumors by learning regionalized propagation patterns and trains to learn the propagation patterns by unsupervised learning is proposed. In addition, a source-enhanced residual graph convolution layer (SRGCL) is designed to improve the graph neural network (GNN) oversmoothness and increase the depth limit of the rumor detection methods-based GNN. Experiments on Twitter15 and Twitter16 show that the proposed model performs better than the baseline approach on rumor detection and early rumor detection.