Detection of Spam Tweets by Using Machine Learning Methods

Abstract

Nowadays, we live in a digital world where we can connect and communicate with anybody all over the world with the help of social networking sites like Twitter and Facebook. Among many other social networking sites, Twitter is a rapidly growing platform. Due to its popularity, it has to face many spammers, which spread forged, bogus, and misleading information. Due to this spamming, it is difficult for researchers to get un-spammed and authentic information from Twitter. In this paper, we introduce an autonomous model that classifies between spammer and legitimate account tweets using machine learning algorithms. We are using API provided by Twitter for the dataset and apply four machine learning classifiers: Decision tree, Support Vector Machine, Naïve Bayes, and Logistic Regression. These machine learning classifiers evaluate on the basis of accuracy, precision, recall and F1-score. The result shows that Naïve Bayes classifier achieve the best result with 96.83 percent accuracy among four classifiers.

Keywords: Spam, Twitter, Machine Learning, Classifiers, Social Networking Site