Suppose that, alternatively, you found the following plot, where the training curve, denoted by $\text{'}CV$ train' on the

plot below, is the result of using $5-$fold cross$-$validation on the training data set only $($NOTE; despite the possibly

confusing label, the $"$CV train" plot refers to the CV plot obtoined during training: that is$,$ it is the average, over the $5$

folds on the training data of the CV test$/$hold$-$out error obtained for each degree value$),$ while the test curve,

denoted by "test" on the plot below, is the result of evaluating on the test data set the classifier that the

regression$-$learning algorithm outputs for the corresponding polynomial$-$degree $d$ using all the training data as

input.