Tackling Misinformation in Mobile Social Networks A BERT- LSTM Approach for Enhancing Digital Literacy

Abstract

The rapid fusion of mobile Internet with the media industry has exponentially accelerated the production and dissemination of misinformation, significantly impacting society. Mobile social networks, in particular, act as fertile grounds for the rapid spread of false news, demanding innovative oversight mechanisms to mitigate this digital epidemic. Our study introduces a robust detection model for false news in mobile social networks, leveraging the synergistic capabilities of Bidirectional Encoder Representations from Transformers (BERT) and Long Short-Term Memory (LSTM) networks. BERT's prowess in contextual word vector extraction, combined with LSTM's sequential data processing strength, provides a nuanced understanding of news content authenticity. We present empirical evidence showcasing the superior performance of our model, which outstrips conventional classifiers like random forest and logistic regression, with an impressive accuracy of 93.51%, recall of 91.96%, and an F1 score of 92.73%. Beyond mere detection, our approach advocates for the empowerment of users, fostering enhanced digital literacy through the development of critical skills necessary to discern credible information. By integrating BERT and LSTM, our model not only effectively flags misinformation but also serves as an educational tool, guiding users towards informed decision-making in the realm of mobile social networks. This research underscores the pivotal role of advanced computational techniques in the fight against misinformation, spotlighting the transformative potential of AI in bolstering digital literacy in an era inundated with ambiguous information.