RT info:eu-repo/semantics/article
T1 How performance metric choice influences individual tree mortality model selection
A1 Vázquez Veloso, Aitor
A1 Núñez Bravo, Andrés
A1 Toraño Caicoya, Astor
A1 Pretzsch, Hans
A1 Bravo Oviedo, Felipe
K1 Recursos forestales
K1 Mortalidad arbórea
K1 Modelos biométricos
K1 Modelización forestal
K1 Supervivencia
K1 Clasificación binaria
K1 Área bajo la curva de precisión-recuerdo (AUCPR)
K1 Coeficiente de correlación de Mathews (MCC)
K1 3106 Ciencia Forestal
K1 3106.08 Silvicultura
AB Understanding tree mortality is crucial to understand forest dynamics and is essential for growth models and simulators. Although factors such as competition, drought, and pathogens drive mortality, their underlying mechanisms remain difficult to model. While substantial attention has focused on selecting appropriate algorithms and covariates, evaluating individual tree mortality models also requires careful selection of performance criteria. This study compares seven different metrics to assess their impact on model evaluation and selection. Results show that candidate models exhibited varying performances across metrics and that the choice of metric significantly influences the selection of the best model. When no confusion matrix was available, the area under the precision-recall curve (AUCPR) emerged as a more reliable alternative to the area under the ROC curve (AUC), offering a more informative assessment for imbalanced datasets. When a confusion matrix was available, Cohen’s Kappa coefficient (K) and Matthews correlation coefficient (MCC) outperformed accuracy-based metrics, providing a fairer evaluation of both live and dead tree classifications. These findings emphasize the importance of choosing appropriate evaluation standards to enhance mortality model assessment and ensure reliable predictions in forestry applications.
PB Springer Nature
SN 1993-0607
YR 2026
FD 2026
LK https://uvadoc.uva.es/handle/10324/82616
UL https://uvadoc.uva.es/handle/10324/82616
LA eng
NO Journal of Forestry Research, enero 2026, vol. 37, n. 49.
NO Producción Científica
DS UVaDOC
RD 28-feb-2026