GoogleCloudAiplatformV1ModelMonitoringObjectiveConfigPredictionDriftDetectionConfig constructor

const GoogleCloudAiplatformV1ModelMonitoringObjectiveConfigPredictionDriftDetectionConfig({ Input<Map<String, String>>? attributionScoreDriftThresholds, Input<GoogleCloudAiplatformV1ThresholdConfig>? defaultDriftThreshold, Input<Map<String, String>>? driftThresholds, })

Creates a new GoogleCloudAiplatformV1ModelMonitoringObjectiveConfigPredictionDriftDetectionConfig. attributionScoreDriftThresholds Key is the feature name and value is the threshold. The threshold here is against attribution score distance between different time windows. defaultDriftThreshold Drift anomaly detection threshold used by all features. When the per-feature thresholds are not set, this field can be used to specify a threshold for all features. driftThresholds Key is the feature name and value is the threshold. If a feature needs to be monitored for drift, a value threshold must be configured for that feature. The threshold here is against feature distribution distance between different time windws.

Implementation

const GoogleCloudAiplatformV1ModelMonitoringObjectiveConfigPredictionDriftDetectionConfig({
  this.attributionScoreDriftThresholds,
  this.defaultDriftThreshold,
  this.driftThresholds,
});