Predictive Analysis of Cyberbullying on X Data using Multi-Model Supervised Techniques
R K Pongiannan, P. U. Poornima, Lourdu Jame S, M Pemila, Hritik Kumar, Abhinav D Trivedi · 2024
Cyberbullying predictive analysis on Twitter data has garnered considerable attention owing to the escalating prevalence of harmful online behavior. This study proposes a novel approach utilizing a multi- model supervised technique to predict instances of cyberbullying on Twitter. The proposed technique aims to bolster the accuracy and effectiveness of cyberbullying detection by amalgamating textual, social, and network features. Twitter data sets containing both cyberbullying and non-cyberbullying instances are used to train and evaluate the models. The textual features include sentiment analysis, bag-of-words, and semantic similarity, while the social features encompass user characteristics, such as follower count and account age. Network features involve analyzing the user's network structure and interaction patterns. Machine learning (ML) algorithms such as random forests (RF), support vector machines (SVM), and neural networks (NN), are utilized to construct and assess models. The Experimental results show that the combined approach achieves improved performance in effective prediction compared with individual models. The study also emphasizes the importance of feature selection in improving model accuracy. By accurately identifying cyberbullying incidents on Twitter, this research contributes to the development of effective strategies and countermeasures to mitigate the harmful effects of cyberbullying.