Shivansh212/Algerian-Forest-Fire

# Algerian Forest Fire Prediction This project uses machine learning to predict the **Fire Weather Index (FWI)** based on meteorological data from two regions in Algeria (Bejaia and Sidi-Bel Abbes). The project is built in two main parts: 1. **Data Cleaning (`clean.ipynb`):** Loads the raw dataset, performs extensive cleaning, feature engineering, and saves the result. 2. **Model Training (`fire.ipynb`):** Loads the cleaned data, applies scaling, trains a Ridge Regression model, and saves the final model and scaler. --- ## Dataset The original dataset is the "Algerian Forest Fires Dataset" from the UCI Machine Learning Repository. It contains meteorological data and fire-related indices. * The **raw dataset** is not included in this repo. * The **cleaned dataset** (`algerian_cleaned_data.csv`) is provided and is the starting point for the modeling notebook. --- ## 📂 Project Structure & File Guide * `clean.ipynb`: Jupyter Notebook for loading the raw data, performing EDA (Exploratory Data Analysis), cleaning, and feature engineering. * `fire.ipynb`: Jupyter Notebook for model training. It loads the clean data, splits it, applies `StandardScaler`, and trains a **Ridge Regression** model. * `algerian_cleaned_data.csv`: The cleaned data, ready for modeling. This is the output of `clean.ipynb`. * `ridge.pkl`: The final, trained **Ridge Regression model** object. * `scaler.pkl`: The saved **StandardScaler** object used to transform the features. * `requirements.txt`: A list of all necessary Python packages. --- ## 🛠️ Tech Stack * Python * Pandas * NumPy * Scikit-learn (for Ridge Regression, StandardScaler, train_test_split) * Matplotlib & Seaborn (for visualization) * Jupyter Notebook