Student Research Assistant (Cyberbullying Detection Deep Learning Project) - Delta State University
Text
ClassificationDeveloped a deep learning cyberbullying detection system by preparing and using an English text dataset for harmful-content classification. Applied common NLP preprocessing steps such as text cleaning, lowercasing, tokenization, and sequence padding to reduce noise and make the data suitable for LSTM training. Performed model training, optimization, and comparative evaluation to assess classification performance and reliability for content moderation use cases.• Prepared Kaggle-sourced text data using cleaning/tokenization/padding pipelines for LSTM modeling.• Trained and optimized an LSTM-based framework for harmful text classification.• Evaluated with accuracy, precision, recall, and F1 to quantify balanced detection performance.• Presented research findings to the departmental group focused on NLP architectures for AI-driven moderation.