kumulant

PageHinkleyResult

@Serializable
@SerialName(value = "PageHinkleyResult")
data class PageHinkleyResult(val delta: Double, val threshold: Double, val count: Long, val mean: Double, val cumulativePositive: Double, val cumulativeNegative: Double, val minPositive: Double, val maxNegative: Double, val alarmUp: Boolean, val alarmDown: Boolean) : Result(source)

Snapshot from a PageHinkleyStat change-point detector.

Constructors

Link copied to clipboard
constructor(delta: Double, threshold: Double, count: Long, mean: Double, cumulativePositive: Double, cumulativeNegative: Double, minPositive: Double, maxNegative: Double, alarmUp: Boolean, alarmDown: Boolean)

Properties

Link copied to clipboard

True when either drift test has fired.

Link copied to clipboard

True when maxNegative - cumulativeNegative > threshold.

Link copied to clipboard

True when cumulativePositive - minPositive > threshold.

Link copied to clipboard
val count: Long

Number of observations folded in.

Link copied to clipboard

Cumulative negative-drift signal m_t = Sum (x_i - mean + delta).

Link copied to clipboard

Cumulative positive-drift signal m_t = Sum (x_i - mean - delta).

Link copied to clipboard

Tolerance applied on each side; positive deviations smaller than this are absorbed.

Link copied to clipboard

Running maximum of cumulativeNegative, used as the baseline for the downward test.

Link copied to clipboard

Running mean of all observations.

Link copied to clipboard

Running minimum of cumulativePositive, used as the baseline for the upward test.

Link copied to clipboard

Alarm threshold; the alarm fires when either drift exceeds this.

PageHinkleyResult

constructor(delta: Double, threshold: Double, count: Long, mean: Double, cumulativePositive: Double, cumulativeNegative: Double, minPositive: Double, maxNegative: Double, alarmUp: Boolean, alarmDown: Boolean)(source)

alarmDown

True when maxNegative - cumulativeNegative > threshold.

alarmUp

True when cumulativePositive - minPositive > threshold.

alarm

True when either drift test has fired.

count

Number of observations folded in.

cumulativeNegative

Cumulative negative-drift signal m_t = Sum (x_i - mean + delta).

cumulativePositive

Cumulative positive-drift signal m_t = Sum (x_i - mean - delta).

delta

Tolerance applied on each side; positive deviations smaller than this are absorbed.

maxNegative

Running maximum of cumulativeNegative, used as the baseline for the downward test.

mean

Running mean of all observations.

minPositive

Running minimum of cumulativePositive, used as the baseline for the upward test.

threshold

Alarm threshold; the alarm fires when either drift exceeds this.