Continuous Attention Mechanism Embedded (CAME) Bi-Directional Long Short-Term Memory Model for Fake News Detection

Abstract

The credible analysis of news on social media due to the fact of spreading unnecessary restlessness and reluctance in the community is a need. Numerous individual or social media marketing entities radiate inauthentic news through online social media. Henceforth, delineating these activities on social media and the apparent identification of delusive content is a challenging task. This work projected a continuous attention-driven memory-based deep learning model to predict the credibility of an article. To exhibit the importance of continuous attention, research work is presented in accretive exaggeration mode. Initially, long short-term memory (LSTM)-based deep learning model has been applied, which is extended by merging the concept of bidirectional LSTM for fake news identification. This research work proposed a continuous attention mechanism embedded (CAME)-bidirectional LSTM model for predicting the nature of news. Result shows the proposed CAME model outperforms the performance as compared to LSTM and the bidirectional LSTM model.