Rumor Detection Based on Knowledge Graph and Graph Attention Network

With the rapid growth of the Internet, people are exposed to a significant increase in the amount of information. However, the dissemination of false news has become rampant, causing adverse effects on society. Rumor detection has emerged as an important and challenging research direction aimed at accurately determining the authenticity of messages. This paper proposes a novel rumor detection method based on a knowledge graph and graph attention networks. The method leverages the attention mechanism of graph attention networks to learn topic-enhanced news representations and extract features related to rumor propagation. It then compares the extracted entity features with the entity features in a knowledge base using an entity comparison network. The proposed method is validated on the publicly available SLN dataset, and experimental results demonstrate a 2% improvement in micro F1 and macro F1 performance metrics compared to state-of-the-art baseline models.