TSA Stationarity Test VI

TSA Stationarity Test Details

This VI performs stationarity estimation on a univariate time series by testing the inversion number according to the following steps:

1. Divides a time series X_t into l subsequences. The mean value of each subsequence forms a time series m₁, m₂,…m_l. The standard deviation value of each subsequence forms a time series s₁, s₂,…s_l.

2. Computes the sum S_m (S_s) of inversion number for the time series m₁, m₂,…, m_l (s₁, s₂,…, s_l).

If X_t is stationary, the statistical value e_me_s satisfies the normal distribution with a mean value of 0 and a standard deviation value of 1.

and

Where m_A is the theoretical mean value of S_m or S_s, which equals , and s_A is the theoretical standard deviation value of S_m or S_s, which equals the following equation:

Given the confidence level a:

• If e_m<N_a/2(0, 1), this time series is mean stationary.
• If e_s<N_a/2(0, 1), this time series is variance stationary.

Examples

Refer to the Series Statistical Analysis VI in the labview\examples\Time Series Analysis\TSAGettingStarted directory for an example of using the TSA Stationarity Test VI.