Abstract:The openness characteristic of social networks facilitates the rapid spread of rumors, necessitating effective methods for detecting and managing the abundance of rumors on social media. Existing studies have primarily focused on improving the accuracy of rumor detection, but often overlook the vital aspects of interpretability and explanation of rumor patterns, limiting their credibility and real-world usability. Additionally, previous works have typically examined only a subset of user features, content, and spreading structure, neglecting the analysis of compound rules. To address these limitations, we propose a novel framework for detecting rumor patterns that emphasize comprehensive feature construction and the explanation of compound rules. Our framework incorporates multi-dimensional features, including user characteristics, post content, and the structure of information propagation. Advanced techniques, including large language models (such as ChatGPT) and graph motif discovery algorithms, are employed for feature construction. By leveraging diverse features, crucial integrated rules identified by Rulefit can investigate the contextually dependent associations among various interrelated rumor factors. We consolidate and analyze seven distinct rumor patterns based on the Credible Early Detection Dataset, deriving valuable insights into the inherent characteristics of rumors. The recognition of rumor patterns empowers social media platforms and fact-checking organizations to develop targeted and explainable interventions that effectively mitigate the spread of rumors and safeguard the integrity of information. These interventions greatly enhance the transparency and trustworthiness of rumor management, fostering a more reliable information ecosystem.

Social Media Rumour Detection Through Graph Attention Networks

Interpretable Graph Neural Network for Social Media Rumor Detection

Detecting the Rumor Patterns Integrating Features of User, Content, and the Spreading Structure.

Rumor Detection Based on Knowledge Graph and Graph Attention Network

Rumor Detection Based on Knowledge Enhancement and Graph Attention Network

Early detection of rumors based on source tweet-word graph attention networks.

Joint learning of structural and textual information on propagation network by graph attention networks for rumor detection

Heterogeneous Graph Attention Networks with Bi-directional Information Propagation for Rumor Detection

Rumor detection on social media using hierarchically aggregated feature via graph neural networks

Rumor Detection with a novel graph neural network approach

Rumour detection using graph neural network and oversampling in benchmark Twitter dataset

Rumour Detection and Analysis on Twitter

Explainable Rumor Detection using Inter and Intra-feature Attention Networks

A Novel Fine-Grained Rumor Detection Algorithm with Attention Mechanism

Graph Convolution-Based Joint Learning of Rumor with Content, User Credibility, Propagation Context, and Cognitive as Well as Emotion Signals

Rumor Detection Based on SAGNN: Simplified Aggregation Graph Neural Networks

Call Attention to Rumors: Deep Attention Based Recurrent Neural Networks for Early Rumor Detection

Federated Graph Attention Network for Rumor Detection

GMRD

Epidemiology-informed Network for Robust Rumor Detection

Continuous-Time Dynamic Graph Networks Integrated with Knowledge Propagation for Social Media Rumor Detection