DetectNet: A Hybrid and Rule-Based and BERT – Powered Cyberbullying Detection System
Abstract
ABSTRACT: In online communication platforms, cyberbullying is a serious issue due to the limits of traditional keyword detection methods and the lack of context in automated systems. In this paper, we present Detect-Net, a cyberbullying detection framework that combines rule-based language analysis with BERT-based deep learning. The system detects both clear and subtle bullying behaviors and sorts them into various categories. Detect-Net also offers clear outputs to improve trust and transparency, making it suitable for real-time moderation and online safety. Key Words: Cyberbullying Detection, Natural Language Processing, BERT, Hybrid Classification, Explainable AI, Text Classification, Social Media Safety.
Fields of Study & Subjects
Hate Speech and Cyberbullying Detection
Sentiment Analysis and Opinion Mining
Spam and Phishing Detection
Publication details
- Journal / source
- INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT
- Volume
- 10
- Issue
- 02
- Pages
- 1–9
- Publication date
- 9 February 2026
- Language
- EN