Assessing performance with cross-validation

K-fold cross-validation:

1. split the data in $K$ parts: use dismo::kfold() or just base R:

k <- 5
random_row_order <- sample(seq_len(nrow(d2)), nrow(d2))
d2$k[random_row_order] <- rep(
  seq_len(k),
  each = ceiling(nrow(d2) / k)
)
head(d2$k, 20)

##  [1] 2 2 3 3 2 3 3 1 3 1 4 2 2 3 2 5 1 4 2 1