Surrogate quality metrics for k-fold cross validation

[FacyMcFace]

Hi there,

I'm performing a 5-fold cross-validation, and all quality metrics are calculated for each fold. However, in the end, the shell only shows the following:

Surrogate quality metrics (2-fold CV) for pq_BP2:
sum_squared 0.356434
mean_squared 0.00742572
root_mean_squared 0.0861726
sum_abs 3.10085
mean_abs 0.0646011
max_abs 0.228022
rsquared nan (n/a for cross-validation)

What exactly are these values? Are they the worst values across all surrogate models, the median, or the mean? I couldn’t find anything in the online help or the user guide PDF that clarifies this.

Also, why is the R-squared value not available? Isn’t it possible to calculate that metric for each surrogate model as well (and then build the mean or something)?

Finally, is there a way to retrieve all quality metrics for every surrogate model used in the cross-validation?

Many thanks!