HumaimaRiaz47/Algerian-Forest-Fire-Prediction-Using-Machine-Learning

A Flask-based web application that predicts the Fire Weather Index (FWI) — an indicator of forest fire risk — using a Ridge Regression model trained on the Algerian Forest Fires dataset. The app allows users to input environmental conditions and get real-time fire risk predictions. # 🌲 Algerian Forest Fire Prediction – End-to-End Machine Learning Project This project predicts the **fire risk percentage** for the Algerian forest regions using environmental and meteorological data. It’s a complete end-to-end pipeline — from **data preprocessing and model training** to a **Flask web application**. --- ## 🚀 Project Overview The **Algerian Forest Fire Dataset** contains environmental attributes like temperature, humidity, and wind speed to help predict forest fire occurrences. Using regression models, this project estimates the likelihood (in %) of a fire occurring based on the given inputs. --- ## 🧠 Machine Learning Pipeline ### 1️⃣ Data Collection The dataset was obtained from the **UCI Machine Learning Repository**. It includes meteorological data from two regions of Algeria — **Bejaia** and **Sidi Bel-Abbès**. ### 2️⃣ Data Preprocessing - Handled missing values - Cleaned and standardized data - Encoded categorical features (`Region`, `Classes`) - Applied feature scaling using `StandardScaler` ### 3️⃣ Exploratory Data Analysis (EDA) - Visualized feature correlations - Identified outliers and trends - Checked multicollinearity between features ### 4️⃣ Feature Engineering - Selected key predictors influencing fire risk: **Temperature, RH, Ws, Rain, FFMC, DMC, ISI, Classes, Region** - Performed feature scaling (standardization) ### 5️⃣ Model Training The following models were trained and evaluated: - 🔹 **Linear Regression** - 🔹 **Lasso Regression** - 🔹 **LassoCV** - 🔹 **Ridge Regression** - 🔹 **RidgeCV** The **Ridge Regression model** gave the best performance and was saved as a `.pkl` file for deployment. ### 6️⃣ Model Serialization Two pickle files were created: - `ridge_model.pkl` → Trained Ridge Regression model - `scaler.pkl` → StandardScaler object for input scaling --- ## 🌐 Flask Web Application The Flask web app allows users to input environmental parameters and get the **predicted fire risk percentage** instantly. ### 🖼️ P …