A Machine Learning Based Approach To Detect Fake News In Social Media

The News available on online resources such as social media applications like facebook, twitter etc. are propagating fast that too in a uncontrolled way and is a matter of high concern in its capability to endure a lot of social and national damage. This damage can prove to be destructive. A lot of researchers are working on detecting this unreliable and falsified information. In this paper, we will be making a short analysis on research related work done till now on fake news detection, using supervised ML algorithms which can catgorize news as true or false by utilising python modules such as sci-kit learn and NLP module for literal analysis. This paper is more focused towards NLP rather than the machine learning domain. This textual analysis process will include processes like feature extraction and vectorization which will be performed using python sci-kit module to carry out feature extraction and pre-processing of text from the dataset, as this library deliver helpful tools like implementation of vectorizer for feature extraction process. Then we will evaluate the performance of different ML algorithms based on several performance factors like accuracy, F1 score, precision and recall. Also, we will work on selecting the best features for the machine learning model to give high precision in detecting fake news. Today the majority of the people have a phone with them and are connected to people around the world through the means of social networking sites. These sites gave everyone the power to keep their opinions but today people are using this power to spread mis-information. This is causing mistrust among people and society. Today fake news is the major reason for the communal riots, government distrust, and a lot more.