This project implements a hybrid SOM + Fuzzy Logic system for forest fire risk prediction using real-world weather and fire data. Instead of binary classification, the model outputs a continuous fire risk score.

• Integrates fire and weather data
• Handles class imbalance using undersampling
• Uses SOM for micro-climate zoning
• Applies fuzzy logic for risk estimation
• Converts risk to predictions via thresholding

• Baseline model struggles due to imbalance
• Fuzzy model improves fire detection (recall)
• Full dataset shows good recall but higher false positives

• Risk modeling is better than binary classification for this problem
• Data balancing improves sensitivity
• SOM provides structure but needs further tuning
• Model prioritizes recall (important for fire detection)

pip install -r requirements.txt python src/data_loader.py python src/balanced.py python src/baseline_model.py

Run SOM + Fuzzy pipeline via neuroFuzzy.ipynb

src/
 ├── data_loader.py
 ├── baseline_model.py
 ├── balanced.py
 ├── som_model.py
 ├── fuzzy_model.py

A practical soft computing approach for fire risk modeling under imbalance. Strong foundation for future improvements.