alk231/Algerian_Forest_Fire

# Algerian Forest Fire Prediction 🌲🔥 A machine learning project that predicts the Fire Weather Index (FWI) based on various meteorological factors. This helps in estimating the risk of forest fires in the Algerian region. ## 🧐 What's This Project? This is an end-to-end machine learning application I built to understand how weather conditions affect forest fire risks. It uses the **Algerian Forest Fires Dataset** to train a Ridge Regression model, which is then served through a Flask web application. The idea is simple: you enter weather details like temperature, humidity, and wind speed, and the model tells you the predicted fire risk score. ## 🛠️ Tech Stack - **Python**: Core language - **Flask**: For the web server and API - **Scikit-learn**: For model training (Ridge Regression) and preprocessing - **Pandas & NumPy**: For data manipulation - **AWS Elastic Beanstalk**: Configuration included for deployment ## 📂 Project Structure ``` Algerian_Forest_Fire/ ├── application.py # Main Flask application ├── requirements.txt # Dependencies ├── .ebextensions/ # AWS Elastic Beanstalk config ├── dataset/ # Raw dataset ├── models/ # Saved pickle files (model & scaler) ├── notebook/ # Jupyter notebook for EDA and training └── templates/ # HTML files for the web interface ``` ## 📊 The Dataset The dataset contains data from two regions in Algeria: - **Bejaia Region** (humid) - **Sidi Bel-Abbes Region** (semi-arid) It includes features like: - **Temperature**: Max temperature in Celsius - **RH**: Relative Humidity % - **Ws**: Wind speed in km/h - **Rain**: Total day in mm - **FFMC**: Fine Fuel Moisture Code - **DMC**: Duff Moisture Code - **ISI**: Initial Spread Index - **Classes**: Fire or No Fire (encoded) - **Region**: 0 for Bejaia, 1 for Sidi Bel-Abbes ## 🚀 How to Run locally 1. **Clone the repo** ```bash git clone github.com cd Algerian_Forest_Fire ``` 2. **Cr …