sklearn/cross-val-score-per-metric
| Rule | cross-val-score-per-metric |
|---|---|
| Pack | scikit-learn |
| Severity | warning |
| Fix | none, edit by hand |
| Library versions | every version |
Each cross_val_score call fits the model on every fold again to compute one more score. Pass every scoring to one cross_validate call, which fits once.
accuracy = cross_val_score(model, X, y, cv=5, scoring="accuracy")precision = cross_val_score(model, X, y, cv=5, scoring="precision")scores = cross_validate(model, X, y, cv=5, scoring=["accuracy", "precision"])accuracy, precision = scores["test_accuracy"], scores["test_precision"]What the rule reports
Section titled “What the rule reports”The rule reports cross_val_score(model, X, y, scoring="accuracy") followed by
cross_val_score(model, X, y, scoring="precision") on the same model, data and cv.
How to fix it
Section titled “How to fix it”xpd has no automatic fix, so make the change by hand.
The rule reports the first call when a later call in the same block passes the same model, data
and cv with another scoring. With LogisticRegression, two scorings in one cross_validate
were 1.60 times faster at 2,000 rows and 1.84 times faster at 20,000 rows, and the scores were bit
equal. Neither reader called one of the 14 held-back reports wrong. Recall was 30 of 30.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
Each cross_val_score call fits $M on every fold again to compute one more score. Pass every scoring to one cross_validate($M, $X, $Y, scoring=[...]), which fits once and measured at least 1.6 times faster for two scorings.Browse the other rules of the scikit-learn pack.