statiotest: Statiotest - Test for Strict Stationarity of a Time Series.

In D-Roberts/statiotest: A Second-Order Cumulant Spectrum Test for Strict Stationarity

Description

The function tests the null hypothesis of strict stationarity of the numerical input vector using a second order cumulant spectrum test, statiotest. The alternatives can be violation via unit root and violations of time invariant second moment of the time series.

Usage

statiotest(x = rnorm(1000), L = round(sqrt(length(x))), idemean = FALSE,
  idetrend = FALSE, itrim = 0)

Arguments

x	A vector.
L	A number.
idemean	A boolean value indicating if the mean is removed.
idetrend	A boolean value indicating if the trend is removed.
itrim	An integer value for percentage each tail trimming.

Details

Reference: Hinich M., Patterson D., and Roberts D., Statiotest. A Second Order Cumulant Spectrum Test for Strict Stationarity in the Frequency Domain, to be published.

Value

A list of statistics, including a p-value. statistic: a vector of values that are tested for normality using the Kolmogorov-Smirnoff test. If normal, then fail to reject null. The elements are the real and imaginary part of the complex valued test statistic calculated using the cumulant spectrum of the data.

p.value:The p-value returned by the K-S test.

alternative: "non-stationary"

method: "statiotest"

data.name: Name of the input vector.

data: Returns a copy of the input vector after preprocessing by demeaning, detrending or trimming.

orig_data: Returns a copy of the input vector.

bifreqs: Returns the tuple of frequencies in the principal domain where the cumulant is calculated based on the sample data and then used to calculate the test statistic.

desc_orig: Returns the numerical summaries of the original data.

Examples

statiotest(rnorm(1000))

D-Roberts/statiotest documentation built on May 6, 2019, 12:55 p.m.