Detecting text anomalies in social networks using different machine learning algorithms

Abstract

Today information is very powerful tool and this information is highly unreliable and wrong information can be very manipulative. It can cause harm to life as well. This in today’s world is called fake news or yellow journalism. Media sometime tries to manipulate and take advantage of innocent viewers by making them believe something that is not real. This has come to publicity recently due to the US presidential election and how media was used to operate the results of the election. Sometime news channel want to garner attention hence they intentionally put this news and as it is our human tendency we always go for the bad apple and this spreads faster than a normal true news would have. The news source got what they needed but they did so by making a fool out of millions of people. So how do we make this problem go away? We train different machine learning algorithms to recognize between fake and real and we know that a machine cannot be subjective and its decision will always be fair and mathematical. Hence in this project we create an API which can tell whether given news is real or fake based of our training data.