Fake News Identification for Web Scrapped Data using Passive Aggressive Classifier

Majority of the people get affected with misleading stories spread through different posts on social media and forward them assuming that it is a fact. Nowadays, Social media is used as a weapon to create havoc in the society by spreading fake news. Such havoc can be controlled by using machine-learning (ML) algorithms. Various methods of ML and deep learning (DL) techniques are used to identify false stories. There is a need for identification and controlling of fake news posts that have increased in alarming rate. Here we use Passive-Aggressive Classifier, Random Forest, Gradient Boosting and Decision Tree for fake news identification. Two datasets, Kaggle fake news dataset and as well as dynamically web scrapped dataset from politifact.com website are used. We achieved 88.66% accuracy using Passive Aggressive Classifier.