Swarm optimized Fake News Detection on Social-media textual content using Deep Learning

In this fast-growing era of technology and social media, the fake news is significant predicament in the real- world settings. Fake news prevention and prediction is a challenging undertaking, and it is becoming a blight on society, necessitating early detection of forthcoming hoaxes. News is supposed to be real and accurate, and the need for information to be true supports the early discovery of false news. Fake news detection, however, still seems to be a pattern identification challenge for computer systems. Real news identification is a crucial issue for improved AI information and is becoming into one of the most significant study areas for the field of intelligent autonomous systems. Motivated by the same, in this research, we aim to empirically assess the performance of swarm optimization on fakes news detection using deep learning on social media textual content. Feature extraction and feature selection are the core components for improved predictions and in this study, we implement RoBerta and Firefly Optimization for the same. Optimized feature vector is being used as input for final classification through supervised deep learning techniques namely Bi-LSTM, CNN and VGGNeT. The results are evaluated on two benchmark datasets namely ISOT and Fakenewsnet. Highest accuracy is obtained by Bi-LSTM.