FeatureStatsAndAnomaly

Stats and Anomaly generated by FeatureMonitorJobs. Anomaly only includes Drift.

Fields

featureId string

feature id.

featureStats

value (  Value 
 
format)

feature stats. e.g. histogram buckets. In the format of tensorflow.metadata.v0.DatasetFeatureStatistics.

distributionDeviation number

Deviation from the current stats to baseline stats. 1. For categorical feature, the distribution distance is calculated by L-inifinity norm. 2. For numerical feature, the distribution distance is calculated by Jensen–Shannon divergence.

driftDetectionThreshold number

This is the threshold used when detecting drifts, which is set in FeatureMonitor.FeatureSelectionConfig.FeatureConfig.drift_threshold

driftDetected boolean

If set to true, indicates current stats is detected as and comparing with baseline stats.

statsTime

string (  Timestamp 
 
format)

The timestamp we take snapshot for feature values to generate stats.

Uses RFC 3339, where generated output will always be Z-normalized and use 0, 3, 6 or 9 fractional digits. Offsets other than "Z" are also accepted. Examples: "2014-10-02T15:01:23Z" , "2014-10-02T15:01:23.045123456Z" or "2014-10-02T15:01:23+05:30" .

featureMonitorJobId

string ( int64 
format)

The id of the FeatureMonitorJob that generated this FeatureStatsAndAnomaly.

featureMonitorId string

The id of the FeatureMonitor that this FeatureStatsAndAnomaly generated according to.

JSON representation
{ "featureId" : string , "featureStats" : value , "distributionDeviation" : number , "driftDetectionThreshold" : number , "driftDetected" : boolean , "statsTime" : string , "featureMonitorJobId" : string , "featureMonitorId" : string }

JSON representation

 { 
 "featureId" 
 : 
 string 
 , 
 "featureStats" 
 : 
 value 
 , 
 "distributionDeviation" 
 : 
 number 
 , 
 "driftDetectionThreshold" 
 : 
 number 
 , 
 "driftDetected" 
 : 
 boolean 
 , 
 "statsTime" 
 : 
 string 
 , 
 "featureMonitorJobId" 
 : 
 string 
 , 
 "featureMonitorId" 
 : 
 string 
 }