One evaluated bucket, kept so that a detector's behaviour can be reviewed after the fact. This is the record that makes threshold tuning possible: you cannot pick a sensible sensitivity from a synthetic fixture, only from what a detector actually said about real traffic. It is also what the admin screen draws. Deliberately flat and small - twelve months of an hourly detector is thousands of these.


Properties

PropertyReturnsDescription
anomalousbooleanWhether this bucket was judged anomalous.
bucketDateDateThe first instant of the bucket this entry records.
directionStringWhich side of normal the observation fell on, as the name of an AnomalyDirection constant.
expecteddoubleThe baseline value the bucket was compared against, the median of the comparable earlier buckets.
observeddoubleThe value that was actually observed in the bucket.
percentOfNormalDoubleThe observed value expressed as a percentage of the baseline, for example 40.0 for a bucket running at forty percent of normal.
scoredoubleThe Iglewicz-Hoaglin modified z-score for this bucket against its baseline.
statusStringThe verdict reached for this bucket, as the name of an AnomalyResult.Status constant.

Methods

getBucketDate()

Returns: Date

The first instant of the bucket this entry records.

getObserved()

Returns: double

The value that was actually observed in the bucket.

getExpected()

Returns: double

The baseline value the bucket was compared against, the median of the comparable earlier buckets.

getScore()

Returns: double

The Iglewicz-Hoaglin modified z-score for this bucket against its baseline.

getPercentOfNormal()

Returns: Double

The observed value expressed as a percentage of the baseline, for example 40.0 for a bucket running at forty percent of normal.

getStatus()

Returns: String

The verdict reached for this bucket, as the name of an AnomalyResult.Status constant.

getDirection()

Returns: String

Which side of normal the observation fell on, as the name of an AnomalyDirection constant.

isAnomalous()

Returns: boolean

Whether this bucket was judged anomalous.

To get full access to the Kademi Hub existing customers can login here, or new customers can register here.