Detection of Cyberbullying in Texts Posted by Users of Social Networks Using Machine Learning
A. A. Zotkina, Alexey Ivanovich Martyshkin · 2024
Despite all the benefits that social networks have brought to the world, they also serve as a favorable environment for the growth of electronic crime. This article explores issues related to the detection of cyberbullying in texts posted by social media users using machine learning. It is noted that the social network VKontakte will be used as the social platform for gathering information in the research. The proposed approach consists of three main stages: pre-processing, feature extraction, and the classification stage. Feature extraction is performed using TF-IDF. Additionally, sentiment analysis is employed alongside TF -IDF to extract the polarity of sentences and add them as features to the feature list. The study utilizes five machine learning classifiers: Support Vector Machine, k-Nearest Neighbors, Logistic Regression, Long Short-Term Memory (LSTM). The article concludes with an assessment of the classifier's accuracy and relevant conclusions are drawn.