Identification of Twitter Bots Based on an Explainable Machine Learning Framework: The US 2020 Elections Case Study

Abstract

Twitter is one of the most popular social networks attracting millions of users, while a considerable proportion of online discourse is captured. It provides a simple usage framework with short messages and an efficient application programming interface (API) enabling the research community to study and analyze several aspects of this social network. However, the Twitter usage simplicity can lead to malicious handling by various bots. The malicious handling phenomenon expands in online discourse, especially during the electoral periods, where except the legitimate bots used for dissemination and communication purposes, the goal is to manipulate the public opinion and the electorate towards a certain direction, specific ideology, or political party. This paper focuses on the design of a novel system for identifying Twitter bots based on labeled Twitter data. To this end, a supervised machine learning (ML) framework is adopted using an Extreme Gradient Boosting (XGBoost) algorithm, where the hyper-parameters are tuned via cross-validation. Our study also deploys Shapley Additive Explanations (SHAP) for explaining the ML model predictions by calculating feature importance, using the game theoretic-based Shapley values. Experimental evaluation on distinct Twitter datasets demonstrate the superiority of our approach, in terms of bot detection accuracy, when compared against a recent state-of-the-art Twitter bot detection method.

Unmasking Fake Social Network Accounts with Explainable Intelligence

“…Detecting fake accounts in social networks is a common task tackled in the literature, using different classification algorithms and datasets from different platforms. Authors of [4]- [11] have worked on datasets collected from Twitter to detect fake or bot accounts. Many account features have been gathered in these datasets.…”

“…Moreover, several AI algorithms have been used to accomplish this task. The authors of [4,[6][7][8][9]11] have used several machine learning algorithms, such as XGBoost (XGB), Random Forest (RF), Support Vector Machine (SVM), AdaBoost, and Logistic Regression (LR). Others used Naive Bayes (NB) [14] and K-Nearest Neighbors (KNN) [12].…”

Discovery and Classification of Twitter Bots

“…Authors in (Pierri, Luceri, and Ferrara 2022) analyze the dynamics of Twitter account creation and suspension during major geopolitical events and show that Twitter's moderation policies varied depending on the type of event and the level of scrutiny from the public and media. Finally, in (Shevtsov et al 2022) they propose an explainable ML pipeline for Twitter bot detection during the 2020 US Presidential Elections using a hybrid approach that combined deep learning and rule-based methods to improve the accuracy of bot detection.…”

“…According to the research community, the usage of manipulation techniques implemented with the use of bot accounts is registered during diverse popular topic discussions. More specifically, studies show that bot accounts are involved in discussions around the 2016 and 2020 US Presidential elections (Golovchenko et al 2020;Badawy, Ferrara, and Lerman 2018;Howard, Kollanyi, and Woolley 2016;Shevtsov et al 2022Shevtsov et al , 2023. Besides the US elections a high bot activity with spreading of misleading information is also detected during election periods (presidential/parliamentary/state) in countries like Germany, Sweden, France, Spain and etc.…”