Ann: Positive Results from Australian Adult Clinical Studies, page-30

I believe the samples are too small to split into separate training and testing sets. We need to deliver results within a reasonable confidence interval. I believe that cross-folding is sufficient for the purposes of this publication.

Regardless of what methodology is used to train/validate the algorithm the medical fraternity will require independent verification. This will be provided by the four independent, double-blind trials being conducted both here and in the US.