A robust method for detecting fake news using both machine and deep learning algorithms

Abstract

Spreading fake news and false information on social media is very common and can be done effortlessly due to the huge number of users of each of the various social media platforms. Another reason for having such a speedy spread of fake news (which makes about 40% of the information published on social media platforms) is the inability of these platform to verify the authenticity of the news before allowing it to be published. This research will use information technology to detect fake news/ false information and change this kind of technology from being the cause of the problem to a tool to solve it. This research provides a method that uses both machine learning (ML) and deep learning (DL) algorithms to detect fake information versus real information and compare the performance of the algorithms. The results of this research indicate that the algorithms that use term frequency inverse document frequency (TF-IDF) have achieved better results than the algorithms that use Word2Vec. Long short-term memory (LSTM) algorithm, however, has achieved the best performance; of 99% accuracy -when using TF-IDF, and 94% -when using Word2Vec.