Multimodal Rumor Detection Method in Social Media Based on Propagation Heterogeneity Graph Modeling

Abstract

The invention discloses a social media multimodal rumor detection method based on propagation heterogeneous graph modeling. In the feature extraction stage, a pre-trained model is used to extract text and image information and a deep learning-based graph convolutional neural network model is captured. Structural information of social media, this method can allow information to spread through the constructed graph network according to the propagation characteristics of social media, so as to obtain richer information, and this method can make full use of limited labeled data and a large amount of unlabeled data, Reduce the resource waste of manual labeling. In the stage of rumor detection, the softmax classifier is used to detect rumors by using the features of network structure information and multi-modal information fusion. Through the method proposed in this patent, rumor detection can be realized automatically, quickly and accurately, thereby reducing the dissemination of false information and untrue remarks and their adverse effects.

Description

technical field [0001] The invention relates to the technical field of cyberspace security, in particular to a social media multimodal rumor detection method based on propagation heterogeneous graph modeling. Background technique [0002] With the development of society, traditional social media has become an important source for users to share information, and social media has an influence that cannot be ignored in information dissemination. But what followed was the viral spread of all kinds of false information. The spread of rumors caused public panic, disrupted social order, affected public opinion, and manipulated the focus of the public, becoming a great social instability factor. Therefore, proposing an effective method for automatically detecting false rumor information is of great significance for maintaining social stability and cyberspace security. [0003] In order to suppress the flood of rumor information in social media, the academic community has proposed r...

AI Technical Summary

Problems solved by technology

However, none of the current detection schemes takes these factors into consideration, so the detection accuracy needs to be improved

Images