Methods for Automated Cyberbullying Detection Driven by Natural Language Processing

Abstract

In today’s digitally linked society, cyberbullying is a severe danger that calls for efficient detection and preventive techniques. A strong multi-tiered approach for identifying and categorizing cyberbullying on various internet platforms is proposed in this study. To achieve high accuracy in differentiating bullying from non-bullying information, the system is trained on extensive and diverse datasets using sophisticated machine learning and natural language processing techniques. It improves inclusivity and dependability by taking language and cultural quirks into consideration. With its flexible design, the system adjusts to new trends in cyberbullying and maintains its efficacy over time. This research lessens the effects of online harassment, promotes the development of more inclusive online communities, and supports the establishment of safer digital environments by facilitating proactive interventions.

Keywords: Accuracy, Contextual, Cyberbullying, Detection systems, Lemmatization, Machine learning, Natural language processing, Precision, Semantic, Social media, Social media communities, Statistical

View PDF