Fake News Detection Using Deep Learning Based Approach

In this era of digital world, peoples are interacting through various online applications like Facebook, Instagram, WhatsApp, Twitter etc. these applications are the medium of online communication but also used as the source of information which generally known as news. There are various users who use these platforms as news platforms to get some news from them. From this, it can be analysed that these platforms are playing major role in creating content on regular basis for the society. to ensure that this content is verified and not contain any fake information which leads to conflicts in the society at different levels. In this paper, the deep learning models have implemented on the ISOT dataset which contains 137176 tweets data collected from twitter and some news data collected from different various sources including the news platforms. To extract features, both TD-IDF and bag of words are used. For generating vectors for words, glove embedding is applied. The deep learning models that have implemented are CNN, LSTM, BiLSTM and hybrid approach of CNN-BiLSTM. The highest accuracy that has achieved in this paper is 83% in both the classifiers i.e., CNN and LSTM.