kritikas11/life-cycle-ml-project

Ridge Regression model predicting Fire Weather Index from weather and fire-index data (Algerian Forest Fires dataset). R²: 0.984, MAE: 0.564. Flask app, deployed with Gunicorn. # Algerian Forest Fire Predictor End-to-end ML lifecycle project predicting forest fire risk (FWI — Fire Weather Index) from weather and regional data, built on the Algerian Forest Fires dataset. Served as a Flask web app. --- ## What it does A user enters weather/fire-index readings through a web form, and a trained Ridge Regression model predicts the Fire Weather Index (FWI) — a measure of forest fire risk. **Input features:** - Temperature - RH (Relative Humidity) - Ws (Wind speed) - Rain - FFMC (Fine Fuel Moisture Code) - DMC (Duff Moisture Code) - ISI (Initial Spread Index) - Classes (fire / not fire, from the source dataset) - Region Inputs are scaled using a pre-fit `StandardScaler` before being passed to the Ridge Regression model. --- ## Tech stack **Backend:** Flask **ML:** scikit-learn (Ridge Regression, StandardScaler) **Data handling:** pandas, numpy **EDA/visualization:** matplotlib, seaborn **Deployment:** Gunicorn, configured for platforms like Render (reads `PORT` from environment) --- ## Project structure ``` . ├── application.py # Flask app — home page + prediction route ├── models/ │ ├── ridge.pkl # Trained Ridge Regression model │ └── scaler.pkl # Fitted StandardScaler ├── noteboooks/ │ ├── 2.0-EDA And FE Algerian Forest Fires.ipynb # Exploratory data analysis + feature engineering │ ├── 3.0-Model Training.ipynb # Model training and evaluation │ └── Algerian_forest_fires_dataset_UPDATE.csv # Source dataset ├── templates/ │ ├── index.html │ └── home.html # Prediction form + results └── requirements.txt ``` --- ## How it works 1. `noteboooks/2.0-EDA And FE Algerian Forest Fires.ipynb` — exploratory analysis and feature engineering on the Algerian Forest Fires dataset. 2. `noteboooks/3.0-Model Training.ipynb` — trains the Ridg …