Classification of Hoax/Non-Hoax News Articles on Social Media using an Effective Deep Neural Network

The internet has become a medium to share information rapidly across the globe. On the other hand, this immense growth of the internet has also led to the sharing of excessive misleading and fake news. There have been many classifiers in the existing literature to classify fraudulent news and legitimate news. Still, they require a lot of data before effective learning, which is present in the most significant cases and it is very challenging to obtain. As deep understanding is one of the emerging technologies, in this paper, we discuss how fake news on the web can be classified using deep learning techniques with testing on several datasets and implementing using several classification models. With our proposed deep learning model, we gained an accuracy of 96.25%. The word embedding techniques have improved the text classification methods. The model architecture for the classification of hoax information has optimized to give better results by reducing the loss value. Since the model has trained on two different real-world datasets, it requires hyperparameters to classify the information.