Mpss1701/Algerian-Forest-Fire-FWI-Prediction

End-to-end Machine Learning project on the Algerian Forest Fires dataset using Linear Regression, Ridge, Lasso, and Elastic Net regression models to predict Fire Weather Index (FWI). ## Overview This project aims to predict the **Fire Weather Index (FWI)** using the Algerian Forest Fires dataset. The workflow includes data preprocessing, exploratory data analysis (EDA), feature selection, model training, hyperparameter tuning, and performance evaluation using multiple regression techniques. ## Problem Statement The Fire Weather Index (FWI) is an important indicator used to assess the risk and intensity of forest fires. Accurate prediction of FWI can help authorities and environmental agencies make informed decisions regarding wildfire prevention and management. ## Dataset The project uses the Algerian Forest Fires dataset, which contains meteorological and fire-related observations collected from different regions of Algeria. ### Features * Temperature * Relative Humidity (RH) * Wind Speed (Ws) * Rain * FFMC * DMC * DC * ISI * BUI * Classes * Region ### Target Variable * Fire Weather Index (FWI) ## Project Workflow ### 1. Data Preprocessing * Handling missing values * Data cleaning * Encoding categorical variables * Feature scaling ### 2. Exploratory Data Analysis (EDA) * Distribution analysis * Correlation analysis * Outlier detection * Feature relationship visualization ### 3. Feature Selection * Correlation-based feature analysis * Removal of redundant features * Multicollinearity assessment ### 4. Model Building The following regression models were implemented and compared: * Linear Regression * Ridge Regression * Lasso Regression * Elastic Net Regression ### 5. Hyperparameter Tuning * Cross-validation using Scikit-Learn * Optimal alpha selection for regularized models ### 6. Model Evaluation Performance was evaluated using: * R² Score * Mean Absolute Error (MAE) * Mean Squared Error (MSE) * Root Mean Squared Error (RMSE) ## Technologies Used * Python * Pandas * NumPy * Matplotlib * Seaborn * Scikit-Learn * Jupyter Notebook / Google Colab ## Key Learnings * Regression modeling and evaluation * Regularization tec …