Automated hybrid Deep Neural Network model for fake news identification and classification in social networks

Abstract

The rapid growth of social media has far-reaching impacts on civilization, traditions, and economics, including both beneficial and unfavourable implications. Since social networking sites have become more frequently utilized for transmitting data, they have also become a gateway for the distribution of fake news for diverse financial and legislative goals. Artificial Intelligence (AI) and Natural Language Processing (NLP) approaches have a lot of ability for academics who wish to design models that can recognize fake news automatically. On the other hand, identifying fake news is a difficult issue because it demands systems that describe the news and then contrast it to the actual news to categorize it as fake. Thus, to overcome this, this paper introduces Hybrid Deep Neural Network Model, in which C-DSSM and Deep CNN models have been utilized. It identifies and classifies fake news using the LIAR dataset. According to experimental results, the proposed model obtained an accuracy of 92.60%, a recall of 92.40%, a precision of 92.50%, and an F1 score of 92.50%. Furthermore, the proposed model is compared to earlier studies for fake news identification using the LIAR dataset, and the proposed model's performance is remarkable. As a result, the proposed hybrid model gives better results in detecting and classifying fake news on social networks.