Using Machine Learning Techniques to Identify Bot Accounts on a Social Network

The purpose of the work was to study the problem of identifying bot accounts in social networks. Bot identification is a classification of user accounts based on the analysis of account parameters and publications. The paper provides an overview of machine learning-based bot detection methods such as decision tree, logistic regression, and naive Bayes classifier. A neural network classification model based on the morphological analysis of user publications has been designed and implemented. Evaluation of the effectiveness and comparative analysis of the models was carried out according to several criteria of effectiveness. The efficiency of the application of the neural network model, which conducts the morphological analysis of publications, is shown, not only from the point of view of the complex coefficient of estimation of the operating time and the classification accuracy, but also from the point of view of the correctness of the classification, under the conditions of humanlike behavior of bots.