Table 2 Mean squared error (MSE) and squared bias (bias2) of the different machine learning and statistical models depending on different resampling strategies, namely, k-fold, environmental, and spatial cross-validation. ran (random forest), xgb (extreme gradient boosting), svm (support vector machine), ann (artificial neural network), glmm (generalized linear mixed models) and gwr (geographically weighted regression). 05 and 15 refer to the models trained with five and fifteen variables, respectively

From: Performance of statistical and machine learning-based methods for predicting biogeographical patterns of fungal productivity in forest ecosystems