🔥 Wildfire Risk Prediction Using Regression ML Model

This project performs end-to-end wildfire risk prediction using multiple regression models based on the Algerian Forest Fires dataset. It covers complete data preprocessing, feature engineering, model training, evaluation, and deployment.

---

📌 Project Overview

Wildfires have become increasingly frequent and intense, affecting ecosystems and economies. In this project, we use real-world data from Algerian forests to build predictive models that estimate Fire Weather Index (FWI), which indicates the potential for wildfire risk.

We build multiple regression models and compare their performance using standard metrics.

---

📊 Dataset Information

• Source: Algerian Forest Fires Dataset (UCI Repository)
• Attributes:
  • Temperature, Relative Humidity, Wind, Rain
  • DC, DMC, FFMC, ISI (fire danger indices)
  • FWI (target variable)
• Target: FWI (Fire Weather Index — numerical value)

---

🧹 Data Preprocessing

• Combined two region-wise datasets into one
• Converted region column to numerical category
• Converted all columns to appropriate data types
• Handled missing values
• Feature scaling using StandardScaler
• Split into train and test sets (80:20)

---

🤖 ML Models Used

• Linear Regression
• Lasso Regression
• Ridge Regression
• ElasticNet Regression
• Decision Tree Regressor
• Random Forest Regressor

---

📈 Evaluation Metrics

• Mean Absolute Error (MAE)
• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)
• R² Score

---

✅ Best Model

After comparing multiple models, Random Forest Regressor achieved the best performance based on R² score and lowest errors.

---

📊 Visualizations

• Correlation Heatmap
• Feature Distribution
• Actual vs Predicted Line Plots
• Residual Plots

(All visuals are included in the notebook for deeper insights.)

---

🧰 Tech Stack

• Python
• Pandas, NumPy
• Scikit-learn
• Seaborn, Matplotlib
• Jupyter Notebook

---

🚀 How to Run Locally

1. Clone this repository

git clone https://github.com/udityamerit/Wildfire-Risk-Prediction-Using-Regression-ML-Model.git
cd Wildfire-Risk-Prediction-Using-Regression-ML-Model

2. Install the required libraries

pip install -r requirements.txt

3. Open the notebook

jupyter notebook End-to-End-ML_Project.ipynb

---

🧠 Future Scope

• Add Streamlit/Flask-based UI for public access
• Use time-series wildfire prediction (e.g. LSTM)
• Integrate satellite imagery using deep learning
• Build a real-time dashboard with geolocation mapping

---

🙋 Author

Uditya Narayan Tiwari 🎓 B.Tech CSE (AI/ML) – VIT Bhopal 🔗 Portfolio 🔗 GitHub 🔗 LinkedIn

---

📄 License

This project is open-sourced under the MIT License.