if negative binomial regression model performance metrics for the testing dataset is better than the training dataset although the difference small during external validation $(1)$ what does it mean $?(2)$ Is it acceptable? $(1)$ what are the steps to conduct $5-$fold cross validation in SPSS$-29($please give hypothetical examples$),$ which performance metrics do I use and how do I use them to draw conclusions?