TF-IDF Weighting to Detect Spammer Accounts on Twitter based on Tweets and Retweet Representation of Tweets

Abstract

Twitter is a social media service that is often used (popular) as a means of communication between users. Twitter's popularity makes spammers spam for personal purposes and gains. Bot spammers are user abuse on Twitter social media. Spammers spread spam repeatedly to other users. This spam is done with the aim of achieving trending topics. Spam activity is carried out by imitating the behavior patterns of real users so that they are not detected as acts of Twitter abuse. in this paper proposed a TF-IDF weighting to detect spammer accounts on Twitter based on tweets and retweet representation of tweets. The purpose of this study is to detect Bot Spammers or Humans using a classification technique using the Naive Bayes algorithm. The best experimental results in the division of 70% training data and 30% test data obtained 92% accuracy with precision and recall of 100% and 87.5%, respectively. This shows that it has successfully detected spammer accounts on Twitter.