A Distinctive Approach for Detecting Fake News using Machine Learning

Abstract

Nowadays, social media platforms have evolved into an esoteric method for audiences to consume information. News spreading through the social network are also a great source of information as well. For the advancement of Internet access, the information consumption pattern is dramatically changing. As a consequence of those, fake news has become one of the prime concerns because of its potentiality to endanger a society in different perspectives as well as has a political and social impact. Because false news causes so much confusion among people, we will train a model to identify all types of fake news in response to public demand. For this respective research, we collect real data from many reliable and reputed online news portals and fake news from unreliable resources. For converting text to vector format Bag of Words is used. Besides TF-IDF is used for extracting the feature and then CNN is used for classification. With an 83.14% accuracy our model can efficiently detect fake and real news. This work paves a path for an easy automatic fake news detection system which will be very helpful for us to prevent spreading the false information and helps to find the truth.