基于BGMA模型社交媒体虚假新闻检测研究

打开文本图片集

中图分类号：TP391.1 文献标志码：A DOI：10. 13705/j. issn.1671-6841.2024092

文章编号：1671-6841（2025）03-0012-07

Research on Fake News Detection in Social Media Based on BGMA Model

WANG Jun1'²，MA Xiaoyue'，FU Hongjing'（1. Institute of Big Data Science， Zhengzhou University of Aeronautics ， Zhengzhou 450015， China;2.Henan Daily，Zhengzhou ，China）

Abstract： In order to identify fake news on social media platforms timely and accurately，a BGMA fake news detection model was constructed.The BGMA model at first used the BERT model to extract the semantic features of the textual content，and then the GAT model was used to capture the complex associations and dynamic changes between user behaviors. Finall，the two features were weighted and fused by introducing a multi-attention mechanism.The results showed that the detection performance of the BGMA model could improves the accuracy by 4.06% on the PolitiFact dataset and 19.73% on the GossipCop dataset compared with the BERT-LSTM model. Compared with the GCNFC model， the accuracy was improved by 10.59% on the PolitiFact dataset and 10.47% on the GossipCop dataset. The practical test result proved that the BGMA model could effectively combine text and user features and achieve better fake news detection results.

Key words： fake news detection;graph attention network ； multi-head attention

0 引言

随着元宇宙技术的发展，各级媒体从智能化、融合化，逐步向智慧化平台发展[1]。（剩余11536字）