Abstract
During the Covid-19 pandemic, the widespread use of social media platforms has facilitated the dissemination of information, fake news, and propaganda, serving as a vital source of self-reported symptoms related to Covid-19. Existing graph-based models, such as Graph Neural Networks (GNNs), have achieved notable success in Natural Language Processing (NLP). However, utilizing GNN-based models for propaganda detection remains challenging because of the challenges related to mining distinct word interactions and storing nonconsecutive and broad contextual data. In this study, we propose a Hierarchical Graph-based Integration Network (H-GIN) designed for detecting propaganda in text within a defined domain using multilabel classification. H-GIN is extracted to build a bi-layer graph inter-intra-channel, such as Residual-driven Enhancement and Processing (RDEP) and Attention-driven Multichannel feature Fusing (ADMF) with suitable labels at two distinct classification levels. First, RDEP procedures facilitate information interactions between distant nodes. Second, by employing these guidelines, ADMF standardizes the Tri-Channels 3-S (sequence, semantic, and syntactic) layer, enabling effective propaganda detection through related and unrelated information propagation of news representations into a classifier from the existing ProText, Qprop, and PTC datasets, thereby ensuring its availability to the public. The H-GIN model demonstrated exceptional performance, achieving an impressive 82% accuracy and surpassing current leading models. Notably, the model's capacity to identify previously unseen examples across diverse openness scenarios at 82% accuracy using the ProText dataset was particularly significant.