Research on a Multimodal Detection Model for False News Based on Multi-text Images
In response to the problem of rampant false news in the Internet environment,the study firstly extracts text and im-age features from false news,and then constructs a false news detection model based on multimodal fusion.The results show that the proposed model has a high detection accuracy of 0.839.The att-RNN model in the Weibo dataset has the highest recall rate for real news and the highest accuracy rate for false news,with values of 0.887 and 0.855,respectively.The MVAE mod-el in the CCF competition dataset has the highest recall rate for false news,with values of 0.737.All other indicators of the proposed model are the highest.The proposed model has significant improvement(p＜0.05)compared to MVAE and att-RNN models,and has stronger clustering and discriminability than MVAE model.In summary,the proposed model can accurately detect false news.
万方数据
维普
