DETECTING BOTS IN TWITTER USING MACHINE LEARNING ALGORITHMS

In this modern world bots are encroaching the software world and tend to contribute almost 50% traffic in the internet .Bots are typically software applications that run on scripts which efficiently fetch, analyze large data at speeds incomprehensible to humans. On social media platforms such as Twitter, bots have gained a lot more liberty to disguise themselves and easily be camouflaged. While bots can be useful source of information, a majority contribute to spam and nuisance over Twitter. In this project we are taking into account the twitter user profile and categorizing it as a Bot/Human profile. We have taken 'Users' JSON and extracted 19 features out of it. On this we did feature engineering to extract the most important features out of it. We used machine learning alogrithms such as Random Forest and XGBoost to categorize a user profile and have obtained a good accuracy of 95%