Fake news detection using Ensemble model

Abstract

Fake news is a piece of information that contains intentional false information. The motivation can be anything, from propaganda to personal benefits. For tackling this, we are building a solution based on Deep Learning, a step above the usual Machine learning approach, using these deep learning algorithms we will detect the accuracy of a piece of information. For this, we are using CNN and LSTM as the base algorithms. Apart from the base algorithms used, we are having used the Ensembling technique via the soft voting method which is a model having its accuracy, thus increasing the overall probability of truth for the given dataset. The model is tested with a huge database from Kaggle and trained by various news sites. This is our advancement to the Machine learning approach. With the use of a better deep learning algorithm, we can promise a better end-user experience with better automated and reliable accurate results.