Enhancing Cyberbullying Detection Through Keyword Filtering: A Comparative Study of ML and DL Approaches

Cyberbullying represents the use of online platforms like social media, gaming platforms, and similar platforms to harass a person, hurt a person, or threaten a person. In this study, keyword filtering is introduced to enhance cyberbullying detection. Various machine learning models, including Logistic Regression, Random Forest, and Support Vector Machine, as well as a deep learning model Convolutional Neural Network, are implemented on original and keyword-labelled datasets. The performance of these models is compared using various evaluation metrics. Our findings demonstrate that the performance of models on the keyword-labeled dataset is significantly higher, with Logistic Regression achieving 98.24%, Random Forest at 99.77%, SVM at 99.67%, and CNN at 99.95% accuracy. This study highlights the effectiveness of keyword filtering features in enhancing the detection of cyberbullying, providing a robust approach to creating safer online environments.