Social Media Fake News Detection using mNB in Blockchain

Nowadays, due to the high usage of social media-based global news, verification and authentication is a very challenging task. Most social media platforms are easily enabled to access news anytime, anywhere over the internet, but it also produces a lot of false news and false information simultaneously. Therefore, in such a case, it is necessary to determine whether available information is genuine, whether it is fake or real. This allows users to make confused and lose the trust of social media. A blockchain-based fake news detection can better handle such problems. The proposed classification algorithm is used to detect fake news in training and testing evolution. Another major objective of this work is to revoke the attackers who update the published news. The blockchain-based decentralized peer-to-peer environment has been used to protect the published data even in a vulnerable environment Various features extraction and selection techniques have been used to generate effective training rules and validate the test classifier accordingly. An extensive experimental analysis demonstrates the classification accuracy of fake news detection on the LIAR dataset. The system achieves 95.20% average accuracy for training as well as testing, which is higher than conventional machine learning algorithms like SVM, ANN, NB etc.