azrapatvi/Algerian-Forest-Fire-Prediction

# 🔥 Algerian Forest Fire Prediction (Machine Learning + Flask Web App) ## 📌 Project Overview This project predicts forest fire risk in Algeria using machine learning and provides an interactive **Flask-based web application** for real-time predictions. It covers the complete ML workflow — data preprocessing, EDA, feature selection, model training, evaluation — and deployment using a modern web interface. --- ## 🌐 Web Application Features * User-friendly prediction form (HTML + Bootstrap) * Takes real-time meteorological inputs * Uses trained ML model for prediction * Displays **Fire Weather Index (FWI)** * Indicates fire risk level based on predicted value --- ## 📂 Project Structure ``` Algerian-Forest-Fire-Prediction/ │ ├── Algerian_forest_fire.ipynb ├── feature_selection_and_model_training.ipynb ├── main.py # Flask backend ├── templates/ │ └── index.html # Frontend UI ├── models.pkl # Trained ML models ├── scaler.pkl # Feature scaler ├── dataset/ (optional) ├── README.md ``` --- ## 🧪 Key Steps Performed ### Machine Learning * Data cleaning & preprocessing * Exploratory Data Analysis (EDA) * Feature engineering & selection * Model training (Linear Regression) * Model evaluation * Model serialization using Pickle ### Deployment * Flask backend (`main.py`) * HTML + Bootstrap frontend * Real-time prediction pipeline --- ## 🛠️ Technologies Used * Python * Google Colab * Pandas, NumPy * Matplotlib, Seaborn * Scikit-learn * Flask * HTML, CSS, Bootstrap --- ## 📊 Dataset The dataset includes environmental and meteorological features: * Temperature * Relative Humidity (RH) * Wind Speed (WS) * Rain * FFMC, DMC, ISI indices * Region Target: Fire Weather Index (FWI) --- ## 📈 Model Output The application predicts **FWI (Fire Weather Index)**: * Low FWI → Low fire risk * High FWI → High fire risk ## ⭐ If you find this project useful, consider giving