OMIS 116 (Using evaluation metrics in
model selection)
For train-validate-test - correct answer ✔✔- Use the respective scikit-learn function call for the specific
score, and feed in
y-test
y-pred
y-proba
as appropriate for the function of interest
For roc_auc_score, when would you input y_test and y_proba? - correct answer ✔✔For the positive
class only
roc_auc - correct answer ✔✔Area under ROC curve
average_precision - correct answer ✔✔area under precision-recall curve
f1, f1_macro, f1_micro, f1_weighted - correct answer ✔✔binary f1 score, weighted variants Regression
r2 - correct answer ✔✔r2 score
model selection)
For train-validate-test - correct answer ✔✔- Use the respective scikit-learn function call for the specific
score, and feed in
y-test
y-pred
y-proba
as appropriate for the function of interest
For roc_auc_score, when would you input y_test and y_proba? - correct answer ✔✔For the positive
class only
roc_auc - correct answer ✔✔Area under ROC curve
average_precision - correct answer ✔✔area under precision-recall curve
f1, f1_macro, f1_micro, f1_weighted - correct answer ✔✔binary f1 score, weighted variants Regression
r2 - correct answer ✔✔r2 score
