Detecting Malicious Twitter Bots Using Machine Learning

Abstract

Twitter is widely used and has become important in the lives of many people in the modern world, including politicians, businesspeople, and the media. Twitter is one of the most widely used social networking sites. It allows users to express their ideas on a variety of topics, such as politics, sports, the financial market, entertainment, and more. It is among the quickest ways to transport information. It has a big impact on people's thinking. On Twitter, the number of individuals hiding their identities for nefarious purposes is increasing. It's critical to identify Twitter bots since they present a danger to other users. As a result, it's critical that actual individuals, not Twitter bots, publish tweets. Spam-related topics are posted by a Twitter bot. Consequently, recognising bots helps distinguish spam communications. Machine learning algorithms analyse features from Twitter accounts to classify individuals as authentic or fake. To ascertain if an account was real or not, we used three machine learning techniques in this study: decision trees, random forests, and multinomial naive bayes. The accuracy and classification performance of the algorithms are compared. About 89% of decisions are made correctly by the Multinomial Naive Bayes technique, 90% by the Random Forest algorithm, and 93% by the Decision Tree algorithm. Consequently, it is evident that decision trees outperform Random Forest and Multinomial Nave Bayes in terms of accuracy.