Stylistic features based Approach for Bot Detection

Abstract

Social bots are computer programs that are created for automating general human activities like generation of messages. The major purposes of bots are resend pictures or messages to other users, giving likes to other user posts, automatic human interaction etc. The rising of bots in social network platforms has led to malicious activities such as content pollution like spammers or malware spreaders, dissemination of misinformation. Most of the researchers focused on the detection of bot accounts in the social media platforms to avoid the damages done to the opinions of users. In this work, a stylistic features based approach is proposed for bot detection. In this approach, a set of stylistic features are extracted from the dataset by analysing the writing styles of both bots and humans. Documents are represented as vectors with a set of stylistic features. Each feature value in the vector representation is determined by using the frequency of a feature in a document. two machine learning algorithms such as Support Vector Machine (SVM) and Random Forest (RF) are used in this experiment to evaluate the performance of proposed approach. The bot detection dataset of PAN 2019 competition is considered as a dataset in this experiment. The experimental results show that the RF algorithm attained best accuracy for bot detection than other efficient approaches in bot detection.