sub1:assertion {
  sub1:comparatorGroup dcterms:description "Different ML/DL architectures compared against each other; comparison of input data configurations (spectral bands, indices, temporal features); validation approaches (cross-validation, independent test sets, spatial holdout); and where available, comparison with traditional remote sensing methods (thresholding, spectral indices)" .
  sub1:interventionGroup dcterms:description "Machine learning and deep learning algorithms applied to Sentinel-2 multispectral imagery for wildfire applications, including convolutional neural networks (CNN, U-Net, ResNet, EfficientNet), random forest, support vector machines, gradient boosting methods, and attention-based architectures. Includes both uni-temporal and bi-temporal approaches, as well as fusion with Sentinel-1 SAR data" .
  sub1:machine-learning-algorithms-for-wildfire-detection pico:comparatorGroup sub1:comparatorGroup ;
    pico:interventionGroup sub1:interventionGroup ;
    pico:outcomeGroup sub1:outcomeGroup ;
    pico:population sub1:population ;
    dcterms:description "What machine learning algorithms have been developed and validated for wildfire detection, risk prediction, and burned area mapping using Sentinel-2 imagery, and what are their reported performance metrics, geographic coverage, and application readiness?" ;
    a pico:PICO , sciencelive:DescriptiveResearchQuestion ;
    rdfs:label "Machine Learning Algorithms for Wildfire Detection and Burned Area Mapping Using Sentinel-2 Imagery: A Systematic Review" .
  sub1:outcomeGroup dcterms:description "Algorithm performance metrics (accuracy, precision, recall, F1-score, IoU, overall accuracy, kappa coefficient), geographic transferability, computational requirements, input data requirements, code and model availability, and operational readiness for wildfire management applications" .
  sub1:population dcterms:description "Geographic regions affected by wildfires globally, with focus on areas where Sentinel-2 multispectral imagery has been applied for wildfire-related studies, including Mediterranean Europe, California, Australia, Canada, and other fire-prone ecosystems" .
}