spectrum.test: Permutation-spectrum test for time-series data

In ts.extend: Stationary Gaussian ARMA Processes and Other Time-Series Utilities

Description

Computes the permutation-spectrum test to detect a periodic signal in a real or complex time-series vector. The null hypothesis for the test is that the values in the time-series vector are independent and identically distributed with no signal, and the alternative hypothesis is that the time-series has at least one remaining periodic signal. The test statistic is the maximum scaled intensity of the observed time-series vector. The p-value for the test is the probability of observing a maximum scaled intensity at least as large as the observed value under the null distribution of exchangeability. The null distribution for the test is simulated by applying random permutations to the observed time-series. The number of simulations is controlled by the sims parameter.

Usage

spectrum.test(x = NULL, sims = 10^6, progress = TRUE)

Arguments

x	A vector of time-series values (must have at least two data points)
sims	Positive integer for the number of simulations to perform in the test
progress	Logical; if TRUE the function uses a progress bar to track its simulations

Examples

data(garma)

#Show the intensity of a time-series vector
spectrum.test(SERIES1, sims = 100)

ts.extend documentation built on Nov. 15, 2020, 1:06 a.m.