kirubel24/Amharic-hate-speech-detector

# Amharic Hate Speech Detection This is a prototype-level hate speech detection system for Amharic language text. The model classifies input text into one of three categories: **hate**, **normal**, or **offensive**. --- ## Project Overview - **Data:** Amharic labeled dataset with text samples categorized as hate, normal, or offensive speech. - **Model:** Multinomial Naive Bayes classifier trained on TF-IDF features extracted from cleaned text. - **Preprocessing:** Text cleaning function handles noise removal and normalization. - **Deployment:** Flask web app exposing a REST API endpoint `/predict` and a simple frontend form for testing predictions. --- ## Features - Clean and preprocess Amharic text data - Encode categorical labels - Train/test split and model evaluation with classification report - Save and load model, vectorizer, and label encoder using `joblib` - Flask backend API for prediction - Interactive web frontend for user input and real-time prediction display --- ## Requirements - Python 3.7+ - Required packages (install via `pip install -r requirements.txt`): ```bash Flask scikit-learn pandas joblib this is how you can run the model on localhost end point cd C:/your-project-folder # 1. Create a virtual environment (Windows) python -m venv venv # 2. Activate the virtual environment (Windows PowerShell) .\venv\Scripts\activate # 3. Upgrade pip (optional but recommended) python -m pip install --upgrade pip # 4. Install required packages pip install flask scikit-learn joblib pandas # 5. Run the Flask app python src/app.py # 6. There is a link that look like Running on 127.0.0.1 (Press CTRL+C to quit) # 7. click ctrl+127.0.0.1 so it runs on default browser