You can indeed do v-fold cross-validation quite easily in GeneXproTools. I describe how this can be done for Regression, Classification and Logistic Regression in the Knowledge Base. Follow any of the links below for a detailed description:
Yes, that’s correct. For the sub-sampling in the Settings Panel you can also choose Random for the Validation Set, but Shuffled also works in this case as you set a total of 75 records for validation and chose a sub-sample of 35. But you could have used the entire validation set (the 75 records) and Random sub-sampling.