Optimizing parameters operator perform differently from validation result

xgdgsc

The optimal parameter given by the "Optimizing parameters" operator , when I apply them to a same "validation" operator as in the "Optimizing" operator. It gives different root mean squared error from the result of the "Optimizing parameters" operator. Why does this happen?

MariusHelf

The X-Validation uses random splits to create the different folds. Try setting a specific local random seed in all X-Validation operators to get the exact same results.

Best regards,
Marius