← 04 Supervised Learning with scikit-learn
Area 04 test · Core ML

Supervised Learning with scikit-learn

20 questions · 70% to pass · no time limit. You see the right answers and why after you submit.

  1. 1Why train a baseline before any real model?
  2. 270% of customers stay. A most-frequent baseline's accuracy is…
  3. 3Linear regression chooses w and b to minimise…
  4. 4Logistic regression outputs…
  5. 5In LogisticRegression, a smaller C means…
  6. 6Coefficients of a logistic regression can be compared as importances when…
  7. 7Gini impurity of a node with 25 churners and 75 stayers is…
  8. 8A decision tree picks a split that…
  9. 9Train AUC 1.00, test AUC 0.59 for a tree. This is…
  10. 10Which change reduces a tree's overfitting?
  11. 11Do decision trees need feature scaling?
  12. 12A random forest reduces variance by…
  13. 13Gradient boosting builds trees…
  14. 14Lowering the boosting learning rate usually requires…
  15. 15Why prefer permutation importance over impurity-based feature_importances_?
  16. 16roc_auc_score should be given…
  17. 17In scikit-learn, parameters learned during fit are…
  18. 18Inside Pipeline([("pre", ...), ("clf", RandomForestClassifier())]), the grid key for max_depth is…
  19. 19Why save the whole Pipeline with joblib, not just the classifier?
  20. 20In the lab the forest's accuracy (0.710) is close to the baseline's (0.700) but its AUC is far higher (0.758 vs 0.500). Why?
0 of 20 answered