Cyberbullying Detection using NLP & ML

Overview

This project develops an automated system for detecting and classifying cyberbullying content on social media platforms using a combination of Natural Language Processing (NLP) and Machine Learning techniques.

Cyberbullying is a growing concern, particularly among young users. Manual moderation cannot scale to the volume of content generated daily. This work aims to build an efficient, accurate classifier that can operate at scale.

Methodology

• Text preprocessing — tokenization, stop-word removal, lemmatization
• Feature extraction — TF-IDF, word embeddings, contextual representations
• Classification models — evaluated multiple ML algorithms including SVM, Random Forest, and neural approaches
• Evaluation — precision, recall, F1-score on benchmark datasets

Key Contributions

• Comparative analysis of multiple NLP feature representations for cyberbullying detection
• Demonstrated effectiveness of combined linguistic and statistical features
• Scalable pipeline suitable for real-time social media monitoring

Publication

Published at the 2024 International Conference on Advances in Computing Research on Science Engineering and Technology (ACROSET), IEEE, Acropolis Institute of Technology & Research, Indore (Sep 27–28, 2024).

(Mankar et al., 2024)

Status

Published — IEEE ACROSET 2024.