siddhi-tawde/Algerian_Forest_Fires_ML

# Algerian Forest Fires Prediction ## Project Overview This project aims to predict the occurrence and severity of forest fires in Algeria based on various environmental attributes. The dataset used is the **Algerian Forest Fires Dataset**, which contains meteorological and fire-related data collected from two regions: **Bejaia (northeast Algeria) and Sidi Bel-abbes (northwest Algeria)**. ## Dataset Information The dataset consists of **244 instances**, with **122 instances for each region**. Data was recorded between **June 2012 and September 2012**. The dataset includes **11 attributes and 1 output attribute (class)**. The instances are classified into **Fire (138 instances)** and **Not Fire (106 instances)** categories. ### Attribute Information 1. **Date**: (DD/MM/YYYY) - Day, Month ('June' to 'September'), Year (2012). 2. **Temp**: Maximum temperature at noon in Celsius (22 to 42°C). 3. **RH**: Relative Humidity (%) (21 to 90). 4. **Ws**: Wind speed in km/h (6 to 29). 5. **Rain**: Total rainfall for the day in mm (0 to 16.8). 6. **FFMC**: Fine Fuel Moisture Code index (28.6 to 92.5). 7. **DMC**: Duff Moisture Code index (1.1 to 65.9). 8. **DC**: Drought Code index (7 to 220.4). 9. **ISI**: Initial Spread Index (0 to 18.5). 10. **BUI**: Buildup Index (1.1 to 68). 11. **FWI**: Fire Weather Index (0 to 31.1). 12. **Class**: Binary classification - **Fire** or **Not Fire**. ## Workflow The project follows a structured machine learning pipeline: 1. **Data Cleaning & Preprocessing** - Handled missing values and inconsistent data entries. - Standardized and formatted the dataset for better analysis. 2. **Exploratory Data Analysis (EDA)** - Visualized data distributions and relationships between features. - Identified important features influencing fire occurrences. 3. **Model Training & Evaluation** - Tested multiple regression models: **Linear Regression, Lasso Regression, Ridge Regression, and ElasticNet**. - Selected **Ridge Regression** as the final model b …