HarshStats/End-2-End-ML-Project-Forest-Fire-Prediction-Deployed-on-AWS

Forest Fire Weather Index (FWI) Prediction App: A Flask web app that predicts FWI for Algerian forests using weather data. Built with a Ridge Regression model trained on the Algerian Forest Fires Dataset (2012). Enter weather details to estimate fire risk instantly. # Forest Fire Weather Index (FWI) Prediction Web App A Flask web application that predicts the Forest Fire Weather Index (FWI) for Algerian forests using weather and environmental data. The app is powered by a Ridge Regression model trained on the Algerian Forest Fires Dataset (2012). Users can input weather details to estimate fire risk instantly. WebApp Link --- ## 📊 Dataset - **Source:** Algerian Forest Fires Dataset (2012) - **Records:** 244 instances (122 each from Bejaia and Sidi-Bel Abbes regions) - **Features:** Weather observations (Temperature, RH, Wind Speed, Rain, etc.), FWI components, Region, and Fire occurrence (Classes) - **Target:** Fire Weather Index (FWI) --- ## 🚀 Project Workflow 1. **Data Cleaning & Preprocessing** - Removed missing values and fixed column names. - Encoded categorical variables (e.g., Classes: 0 = Not Fire, 1 = Fire). - Added region codes (0 = Bejaia, 1 = Sidi-Bel Abbes). 2. **Feature Engineering** - Dropped unnecessary columns (day, month, year). - Checked and removed highly correlated features to reduce multicollinearity. - Standardized features using `StandardScaler`. 3. **Model Training & Evaluation** - Trained and compared several regression models: - Linear Regression - Lasso Regression (with/without cross-validation) - Ridge Regression (with/without cross-validation) - ElasticNet Regression (with/without cross-validation) - Evaluated using Mean Absolute Error (MAE) and R² Score. 4. **Model Selection** - Ridge Regression and ElasticNet performed best and were saved for deployment. 5. **Web App Deployment** - Built a Flask web app for real-time FWI prediction. --- ## 📈 Model Comparison | Model | MAE | R² Score | |-----------------|----------|-----------| | Linear Regression | 0.5468 | 0.9848 | | Lasso | 1.1332 | 0.9492 | | LassoCV | 0.6200 | 0.9821 | | Ridge | 0.5642 | 0.9843 | | RidgeCV | 0.5642 | 0.9843 | | ElasticNet …