snowflake.ml.modeling.metrics.confusion_matrix¶

snowflake.ml.modeling.metrics.confusion_matrix(*, df: DataFrame, y_true_col_name: str, y_pred_col_name: str, labels: Optional[Union[_SupportsArray[dtype], _NestedSequence[_SupportsArray[dtype]], bool, int, float, complex, str, bytes, _NestedSequence[Union[bool, int, float, complex, str, bytes]]]] = None, sample_weight_col_name: Optional[str] = None, normalize: Optional[str] = None) → Union[ndarray[Any, dtype[int64]], ndarray[Any, dtype[float64]]]¶

Compute confusion matrix to evaluate the accuracy of a classification.

By definition a confusion matrix $C$ is such that $C_{i, j}$ is equal to the number of observations known to be in group $i$ and predicted to be in group $j$ .

Thus in binary classification, the count of true negatives is $C_{0,0}$ , false negatives is $C_{1,0}$ , true positives is $C_{1,1}$ and false positives is $C_{0,1}$ .

Parameters:

df – snowpark.DataFrame Input dataframe.
y_true_col_name – string or list of strings Column name representing actual values.
y_pred_col_name – string or list of strings Column name representing predicted values.
labels – list of labels, default=None List of labels to index the matrix. This may be used to reorder or select a subset of labels. If None is given, those that appear at least once in the y true or y pred column are used in sorted order.
sample_weight_col_name – string, default=None Column name representing sample weights.
normalize – {‘true’, ‘pred’, ‘all’}, default=None Normalizes confusion matrix over the true (rows), predicted (columns) conditions or all the population. If None, confusion matrix will not be normalized.

Returns:

ndarray of shape (n_classes, n_classes): Confusion matrix whose i-th row and j-th column entry indicates the number of samples with true label being i-th class and predicted label being j-th class.

Return type:

Raises:

ValueError – The given labels is empty.
ValueError – No label specified in the given labels is in the y true column.
ValueError – normalize is not one of {‘true’, ‘pred’, ‘all’, None}.