What explains the difference in the code to plot the validation curve in the lecture notebook to that in quiz notebook. Both the models are classifier model so I am assuming the template code should have been same, except a few arguments.
Please correct me if I am wrong. Thanks.
On the lecture notebook Overfit-generalization-underfit we are calling the validation_curve
with scoring="neg_mean_absolute_error"
, as the DecisionTreeRegressor
is a regression.
In the case of the Wrap-up quiz 2, Q7 asks you to use scoring=balanced_accuracy
which is in fact used for classification problems.
You may also want to give a look at the documentation on available metrics here.
I hope this answers your question.