Wildfire Ignition Point Classification

Overview

This project aims to build a classification model to predict whether a given geographical point has a high probability of being an ignition point for a wildfire.

The dataset includes various features related to vegetation, weather conditions, and geographical information for each point. The target variable (ignition) indicates whether the point is a known ignition point (1) or not (0).

Dataset

The dataset contains various features, including:

• Geographical Information: elevation, slope, aspect
• Vegetation Class: cropland, forest, wetland, etc.
• Weather Data: max_temp, avg_temp, max_wind_vel, avg_rel_hum, sum_prec
• Population: pop_dens
• Target: ignition (1 if the point is an ignition point, 0 otherwise)

Each row in the dataset represents either an ignition point or a non-ignition point, along with the associated environmental and weather conditions.

Project Structure

This project is organized as follows:

├── data                    # Data folder (dataset not included)
├── src                      # Source code for the project
│   ├── data                 # Data loading and preprocessing scripts
│   │   ├── data_cleaning.py
│   │   ├── data_processing.py
│   │   └── load_data.py
│   ├── features             # Feature engineering scripts
│   │   └── feature_engineering.py
│   ├── models               # Model training and evaluation scripts
│   │   ├── evaluate_model.py
│   │   └── model_training.py
│   └── visualization        # Scripts for visualizing results
│       └── visualize.py
├── README.md                # Project overview
└── requirements.txt         # Dependencies