SignalFx Developers Guide

kpss()

Calculates the Kwiatkowski–Phillips–Schmidt–Shin (KPSS) statistic on the specified time window of the stream. You can’t use it to aggregate time series. To learn more about KPSS, see https://en.wikipedia.org/wiki/KPSS_test.

Syntax

kpss(over=duration[, mode=calcmode])

Table 1. Parameter definitions
Parameter Type Description

duration

Duration (number and duration units indicator)
Durations are specified as a number followed by a single character:

  • s: seconds

  • m: minutes

  • h: hours

  • d: days

Default is None.

Required: The duration over which to calculate the KPSS statistic of each input timeseries

calcmode

string. Default is "level"

Optional: Whether to test for stationary ("level") or trend-stationary ("trend") KPSS statistics

Returns a reference to the input stream object

Notes

Calculates the KPSS statistic for the null hypothesis that the time series is stationary (if mode is 'level') or trend-stationary (if mode is 'trend'). The method doesn’t use lags to estimate the variance of the errors of the model, so it assumes that errors are independently and identically distributed.

Examples

1
2
A = data('jvm.cpu.load')
A.kpss(over='1h').publish('1h_kpss_stat')

© Copyright 2019 SignalFx.

Third-party license information