Efficient Cyberbullying Detection on Social Media Using Lightweight Transformer and BiLSTM Hybrid Architecture

Koritala, Anusha (2025) Efficient Cyberbullying Detection on Social Media Using Lightweight Transformer and BiLSTM Hybrid Architecture. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration Manual) Download (748kB) | Preview

Abstract

Accurate detection of cyberbullying on social media platforms is essential for safeguarding online communities, enabling timely intervention, and supporting healthier digital interactions. While traditional machine learning and deep learning models have achieved progress, they often struggle to capture nuanced contextual meanings, slang, and sequential dependencies present in short, informal online messages, while also maintaining computational efficiency for real-time moderation. This study proposes a lightweight hybrid architecture combining DistilBERT for context-rich transformer embeddings with a BiLSTM layer to model bidirectional sequential dependencies. DistilBERT efficiently encodes subtle linguistic cues with reduced computational overhead, while the BiLSTM refines temporal patterns to improve classification accuracy and robustness. Experimental evaluations on a benchmark cyberbullying dataset demonstrate that the leveraged model outperforms baseline approaches such as Random Forest, LightGBM, KNN, standalone BiLSTM, and CNN models, achieving superior precision, recall, and F1-scores. The results validate the hybrid DistilBERT–BiLSTM model’s capability to deliver high accuracy, interpretability, and lightweight deployment, making it an effective and scalable solution for real-time cyberbullying detection in social media environments.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Yalemisew, Abgaz UNSPECIFIED
Subjects:	B Philosophy. Psychology. Religion > Psychology > Aggressiveness > Bullying Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4150 Computer Network Resources > The Internet > World Wide Web > Websites > Online social networks T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > The Internet > World Wide Web > Websites > Online social networks
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	25 Aug 2026 15:11
Last Modified:	25 Aug 2026 15:11
URI:	https://norma.ncirl.ie/id/eprint/9636

Actions (login required)

View Item