dataeval.detectors.drift¶
Drift detectors identify if the statistical properties of the data has changed.
Submodules¶
Update strategies inform how the drift detector classes update the reference data when monitoring. |
Classes¶
Drift detector using the Cramér-von Mises (CVM) Test. |
|
Drift detector employing the Kolmogorov-Smirnov (KS) distribution test. |
|
Drift detector using Maximum Mean Discrepancy (MMD) Drift Detection with permutation test. |
|
Multivariant Domain Classifier |
|
Drift detector using model prediction uncertainty. |
|
Protocol for reference dataset update strategy for drift detectors |
Output Classes¶
Output class for |
|
Class wrapping the results of the classifier for drift detection and providing plotting functionality. |
|
Output class for univariate drift detectors. |