Hate Speech Detection System 🗣️🚫

Project Type: Group Project

Technologies Used: Python, TensorFlow/Keras, NLP, LSTM, Bidirectional LSTM, GloVe Embeddings, Pandas, Matplotlib, Seaborn, SklearnDataset: HateSpeechDatasetBalanced.csv

Overview

The goal of this project was to build a Hate Speech Detection System to classify text into three categories: hate speech, offensive language, and non-offensive. Each team member trained a different model, and the best-performing model was selected. I developed and optimized a Bidirectional LSTM model, which achieved the highest accuracy and was chosen for presentation.

Key Contributions

✅ Model Development: Developed a Bidirectional LSTM using TensorFlow/Keras for text classification.

✅ Data Preprocessing: Cleaned and tokenized text data, applied GloVe embeddings, and padded sequences for model input.

✅ Hyperparameter Tuning: Experimented with different LSTM layers, dropout rates, and optimizers to boost performance.

✅ Evaluation & Comparison: Achieved the highest accuracy (89.05%) and validation accuracy (87.30%) among all models tested.

✅ Visualization: Created confusion matrices and classification reports for performance analysis.

Results

📌 Final Model Accuracy: 89.05%📌 Validation Accuracy: 87.30%📌 Loss: 0.1660 (train), 0.1973 (validation)📌 Confusion Matrix: Showed strong classification performance across all classes.

🔗 Interactivity & Links

​📌GitHub Repository (https://github.com/mqarmout/hate_speech_detection_system.git)