Comparative Analysis of Machine Learning Algorithms and Datasets for Detecting Cyberbullying on Social Media Platforms
Palagati Anusha, Santhosh Kumar Balan, S. Arun Joe Babulo, Laxmi Raja, K.K. Natarajan, K. Rajkumar · 2024
Cyberbullying is defined as sharing of undesirable material in the cyberspace, primarily- the use of the social sites that instigate hostility and create feelings of hatred among the individuals. Another increasing threat due to increase use of internet-based communication is cyberbullying especially on the site such as Twitter. Different ML and NLP approaches including support vector machine (SVM), Naϊve bayes, Random Forest, and logistic regression is employed to detect cyberbul Bullivating. These methods for supervised learning mine patterns in text and image, allowing computers to promptly flag risky interactions. Some of these approaches also utilizes OCR for the identification of image-based bullying. Consequently, the objective of this research is to identify the correct algorithm and dataset for improving the identification of cyberbullying and solving the problem that manifests itself in the steady increase in cases of this type of bullying.