The package allows creating spatially or environmentally separated folds for cross-validation to provide a robust error estimation in spatially structured environments. Moreover, it permits to investigate and visualise the effective range of spatial autocorrelation in continuous raster covariates and point samples to find an initial realistic distance band to separate training and testing datasets spatially described in Valavi, R. et al. (2019), https://doi.org/10.1111/2041-210X.13107.

The package runs ecological niche models over all combinations of user-defined settings (i.e., tuning), performs cross validation to evaluate models, and returns data tables to aid in selection of optimal model settings that balance goodness-of-fit and model complexity. Moreover, it provides functions to partition data spatially (or not) for cross validation, to plot multiple visualizations of results, to run null models to estimate significance and effect sizes of performance metrics, and to calculate niche overlap between model predictions, among others. The package was originally built for Maxent models, but the current version allows possible extensions for any modeling algorithm.