A Novel Characterization of the Fake News in Twitter Networks

Nowadays, mobile usage is widely expanding with the help of the internet. This statement also spreads fake news easily and negatively impacts society. As a result, it's critical to create a basic mathematical model to comprehend fake news on social networks. To address this, it develops the auto-regressive integrated moving average (ARIMA) and the long-term short memory (LSTM) to detect false news more accurately to maintain high-level security in social networks. The proposed novel deep learning model accurately identifies fake news based on standard historical data. The performance of this model is validated with three standard metrics: accuracy, sensitivity, specificity, and F1-Score. The proposed LSTM predictor achieved nearly 97.5% accuracy, 94.6% the F1-score, and 91.2
^%
sensitivity on average. The results are compared with other existing methodologies, and the proposed framework outperformed existing methodologies. This framework provides promising solutions to detect and predict fake news.