Table 1.

Details of the parametrization used to tune each of the final classification models.

Classifier	Parameters
kNN	Number of neighbors: 3; Metric: Euclidean; Weight: Distance.
Decision tree	Limit the tree depth: 100; Do not split subsets smaller than: 2; Min. number of instances in leaves 3.
SVM	C: 15; Kernel: Radial Basis Function (RBF); g: auto.
Random forests	Number of trees: 15; Do not split subsets smaller than: 5.
ANN	Neurons in hidden layers: 300; activation: Rectified Linear Unit (Relu); solver: Adam; regularization: 0.02.
Naive Bayes	Non-applicable
AdaBoost	Number of estimators: 80; learning rate: 0,7; classification algorithm: SAMME.R; Regression loss function: Square.

kNN: k-nearest neighbors; SVM: support vector machine; ANN: artificial neural networks.