Detecting Suspicious Activities of Digital Trolls During the Political Crisis

Abstract

With the huge widespread and usage of social media in our life, it becomes an effective tool to share our feeling, opinions and even our political orientations when there is civil unrest or protests. As a result, there are many tries from different formal and informal groups to affect on or reply to any political tweets or trends that rise on social media. The main purpose of these groups is to influence the general opinion to fit the intended goal by spreading fake posts, news and even hate speech. This paper introduces a set of first stage analyses automated methods that can be used to detect the effect of such groups called ”digital trolls”.We considered the recently occurred Iraq unrest and Iraqi people protest as a case study to analyze the activities of Twitter users and detect if there are any external groups try to influence the people’s opinions and orientations during the crisis. We gathered a new related dataset from Twitter that includes tweets and users’ information collected during the crisis. Using this dataset, we analyzed the behavior and activities of twitter users by employing different features and tools. The results show that there are suspicious activities from external groups try to affect the general and normal political orientations in Iraq and there is a high possibility to be a kind of digital trolls. Additionally,our empirical study shows that ”user favourites”, ”tweet favorited”, and ”is a retweet” features can be used effectively to detect the digital trolls in Twitter.