Twitter Bot Detection using Machine Learning Algorithms

In nowadays world lots of people like a businessman’s, Media, politicians, etc., uses Twitter daily & have become an important part of life. Twitter is one of the favourite social networking sites which let the individuals to express their sentiment on various topics like politics, sports, stock market, entertainment etc. It is one of the quickest means of transmission information. It extremely affects people’s viewpoint. An increasing number of people on twitter but hide their identity for malignant purpose. It is dangerous for other users hence the necessity for identifying the twitter bots. So it is essential that tweets are sent by authentic users and not by twitter bots. A twitter bot transmits spam subject matters. Therefore detecting of bots helps to determine spam messages.The characteristics of twitter accounts are utilized as Features in machine learning algorithms to label users as genuine or fake. In this paper, we used three machine learning algorithms to detect the account is fake or real, which are Decision Tree, Random Forest, and Multinomial Naive Bayes The classification performance of the algorithms is compared with their accuracy. The accuracy given by the Decision tree algorithm is 93%, the Random Forest algorithm is 90% and the Multinomial Naive Bayes is 89%. Hence it is seen that the Decision tree gives more accuracy as compared to Random Forest and Multinomial Naive Bayes.