WhiteML-Box: An open-source graphical toolbox for spectral machine-learning-based vegetation trait modelling and mapping

Résumé fourni par la source

WhiteML-Box is an open-source graphical research application for spectral machine-learning-based vegetation trait modelling and raster mapping. It integrates spectral-data import, preprocessing, outlier screening, regression modelling, validation, diagnostic visualisation, model export, and prediction of vegetation traits from compatible multiband GeoTIFF imagery. Version 0.8.0 adds automated scientific and data-integrity tests, leakage-safe nested PLSR component selection, training-only model selection for held-out evaluation, reproducible validation outputs, and continuous integration on Ubuntu and Windows with Python 3.10 and 3.12. The included reference analysis verifies deterministic software behaviour in the recorded environment. It does not establish ecological transferability to other datasets, sensors, vegetation traits, or study regions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.