hemin001/Algerian-Forest-Fire

# 🔥 Algerian Forest Fires Analysis & Prediction 🌲 ### 📘 **Project Overview** This project explores the **Algerian Forest Fires Dataset**, which contains weather and fire data from two regions in Algeria (**Bejaia** and **Sidi Bel-abbes**) during the summer of 2012. The goal is to analyze fire patterns, visualize key features, and build regression models to predict the **Fire Weather Index (FWI)**. --- ## 📂 **Dataset Details** - **Instances:** 244 (122 per region) - **Time Period:** June to September 2012 - **Attributes:** 11 features + 1 class label (Fire/Not Fire) --- ## 🛠️ **What I Did** ### 1. **Data Preprocessing & Cleaning** - Handled missing values and corrected data types - Added a **Region** column (`0` for Bejaia, `1` for Sidi Bel-abbes) ### 2. **Exploratory Data Analysis (EDA)** - Visualized data with **density plots**, **pie charts**, and **boxplots** - Analyzed feature relationships with **heatmaps** and **correlation matrices** ### 3. **Monthly Fire Analysis** - Investigated fire occurrence patterns across months for both regions ### 4. **Feature Scaling** - Applied **Standard Scaling** to normalize features ### 5. **Regression Modeling** - Built models to predict **FWI** using: - **Linear Regression** - **Ridge Regression** - **Lasso Regression** - **ElasticNet Regression** ### 6. **Model Evaluation** - Compared models using: - **Mean Absolute Error (MAE)** - **R² Score** - Performed **Cross-Validation** for Ridge, Lasso, and ElasticNet --- ## 🚀 **Results & Insights** - **Ridge and Lasso regression provided the best predictions for FWI** - **Scaling features improved model accuracy and stability** --- ## 📊 **Technologies Used** - **Google Colab** for development - **Python** (NumPy, Pandas) for data manipulation - **Matplotlib, Seaborn** for visualizations - **Scikit-learn** for machine learning models --- ## 🧠 **Key Learnings** - Importance of **EDA** and **data visualization** - How **scaling** influences regression models - Using ** …