ML.EVAL.CLASSIFICATION.ROC_AUC¶
Returns the ROC AUC classification score.
Syntax¶
ML.EVAL.CLASSIFICATION.ROC_AUC(y_true, y_score, average, sample_weight, max_fpr, multi_class, labels)
Arguments¶
| Name | Type | Default | Description |
|---|---|---|---|
| y_true | object | DataFrame or array object of ground-truth target values. | |
| y_score | object | DataFrame or array object of predicted probabilities or decision scores (e.g. output of ML.PREDICT_PROBA or ML.PREDICT). | |
| average | Any | "macro" | How to average across classes. One of 'micro', 'macro', 'weighted', 'samples', or leave blank for per-class scores. |
| sample_weight | Any | None | Optional DataFrame or array object of per-sample weights. Omit for uniform weights. |
| max_fpr | Any | None | If set, compute the partial AUC up to this false-positive rate (value between 0 and 1). Binary classification only. |
| multi_class | Any | "raise" | Strategy for multiclass classification. 'ovr' = one-vs-rest, 'ovo' = one-vs-one, 'raise' = error on multiclass input. |
| labels | object | None | Optional array of class labels to include in the score, in the order they should appear. |
Examples¶
Examples coming soon
Working Excel formula examples for this function are not yet written.
See also¶
- ML.EVAL.CLASSIFICATION.ACCURACY
- ML.EVAL.CLASSIFICATION.AVERAGE_PRECISION
- ML.EVAL.CLASSIFICATION.BALANCED_ACCURACY
- ML.EVAL.CLASSIFICATION.BRIER_SCORE_LOSS
- ML.EVAL.CLASSIFICATION.D2_LOG_LOSS_SCORE
- ML.EVAL.CLASSIFICATION.F1
- ML.EVAL.CLASSIFICATION.JACCARD
- ML.EVAL.CLASSIFICATION.LOG_LOSS
- ML.EVAL.CLASSIFICATION.MATTHEWS_CORRCOEF
- ML.EVAL.CLASSIFICATION.PRECISION
- ML.EVAL.CLASSIFICATION.RECALL
- ML.EVAL.CLASSIFICATION.TOP_K_ACCURACY