shroud07/Algerian_Forest_FWI_Predictor

# 🌲 Algerian Forest Fire FWI Predictor A machine learning web application built with **Flask** that predicts the **Fire Weather Index (FWI)** of Algerian forests based on meteorological data. The underlying prediction engine utilizes a trained **Ridge Regression** model to deliver robust performance against multi-collinearity. --- ## 📌 Project Architecture & Workflow The application accepts weather observations, normalizes the inputs, and serves predictions via a user-friendly interface. 1. **Exploratory Data Analysis (EDA):** Performed extensive data cleaning, visualization (using `seaborn` and `matplotlib`), and feature selection on the Algerian Forest Fire dataset. 2. **Feature Scaling:** Inputs are scaled using a pre-trained `StandardScaler` pipeline to ensure numeric equilibrium across features. 3. **Model Prediction:** A regularized **Ridge Regression** model evaluates the inputs and predicts the continuous FWI value. 4. **Web Interface:** A Flask application serves a responsive form where users submit custom features and instantly receive predictions. --- ## 📊 Dataset Overview & Features The model is trained on observations covering two main regions of Algeria (**Bejaia** and **Sidi Bel-Abbes**). The feature inputs expected by the interface are: | Feature Name | Description | | :--- | :--- | | **Temperature** | Max temperature in Noon (°C) | | **RH** | Relative Humidity (%) | | **Ws** | Wind Speed (km/h) | | **Rain** | Total daily rainfall (mm) | | **FFMC** | Fine Fuel Moisture Code index | | **DMC** | Duff Moisture Code index | | **ISI** | Initial Spread Index | | **Classes** | Forest fire state classification encoded as numerical values | | **Region** | Numerical encoding for the target region | --- ## 🛠️ Tech Stack & Requirements The project leverages standard data science libraries and a micro web framework: * **Backend Framework:** `Flask` * **Data Processing & Math:** `numpy`, `pandas` * **Machine Learning Framework:** `scikit-learn` * …