Baru Validator
QGIS plugin for validating machine learning models for Baru detection. Supports raster and vector classification inputs, calculates Kappa, QADI, F1-Score, MCC, and other advanced metrics for model quality assessment.
Baru Validator is a comprehensive QGIS plugin designed for validating machine learning models used in Baru detection across satellite imagery (Sentinel 10m, Sentinel 1m, Google Earth). The plugin compares classified raster/vector data against validation datasets (points or polygons) and provides advanced accuracy metrics including Cohen's Kappa, QADI index, F1-Score, Matthews Correlation Coefficient (MCC), Producer's Accuracy, User's Accuracy, and confusion matrices. Ideal for assessing model generalization when replicating trained models across different geographic areas without collecting new training samples.