FAKE NEWS AND HATE SPEECH DETECTION WITH MACHINE LEARNING AND NLP

Abstract

With the increase in the ease of publishing and distributing news over the years, the fake
news and hate speech propaganda has taken up a huge chunk in our daily routine, whether we
can identify them or not. We must act to tackle this scenario as these can have contributed to the
increase in political or communal hatred, which can cause severe damage to society. In the
proposed research, we are going to extract features of the language and content by collecting
examples of both real and fake news. We are going to train a model to classify fake news articles
based on the NLP technique called TD-IDF (term frequency-inverse document frequency)
vectorization which gives us the importance of each keyword within the news article or the
speech. Then with the application of logistic regression we can classify the article/speech that
can help us to identify such articles and deal with them.