Bengali Fake News Detection

Abstract

Yellow journalism has become a buzzword for everyone nowadays. Increasing use of internet and social media makes people more vulnerable to fake news. To gain popularity and to have profit through clickbait news publisher and social media circulate fake news to deceive people by creating interesting content of a specific topic. The spread of falsified news has become severe in recent times throughout the world. Though recently some existing system is made to classify and to detect fake news for English news article, not much work has been reported for Bengali news. In this work, we consider Bengali fake news classification considering South Asian Context. More than 200 million people speak Bengali and their way of communication is Bengali. In our Bengali fake news classification system, data mining algorithm is used to classify fake and real news. We have also introduced web interface based on our classifier to check whether a news article written in Bengali language fake or real. The classification model has 85% accuracy with random forest classifier.