As we use k-fold, we further need not evaluate the model performance on a holdout dataset as in a single train-test split process.
a. True
b. False