MAR_Variance: A funtion to generate a multivariate autoregressive process...

In SNSeg: Self-Normalization(SN) Based Change-Point Estimation for Time Series

A funtion to generate a multivariate autoregressive process (MAR) model in time series for testing change points based on variance and autocovariance

Description

The function MAR_Variance is used for generating MAR model(s) for examples of the functions SNSeg_Uni, SNSeg_Multi, and SNSeg_HD.

Usage

MAR_Variance(reptime, type = "V3")

Arguments

reptime

The number of time series to be generated

type

The type of time series for simulation, which includes V1, V2, V3 , A1, A2 and A3. The V-beginnings are for testing the variance, and the A-beginnings are for testing the autocorrelation. The simulated time series come from supplement of Zhao et al. (2022) doi:10.1111/rssb.12552. Default type is V3. The time length and "true change-points locations" (cps) for each type are as follows: V1: cps at 400 and 750 with a time length of 1024. V2: cps at 125, 532 and 704 with a time length of 1024. V3: cps at 512 and 768 with a time length of 1024. A1: cps at 400 and 750 with a time length of 1024. A2: cps at 50 with a time length of 1024. A3: cps at 512 and 768 with a time length of 1024.

Value

Returns a matrix of the simulated MAR processes. The number of columns of this matrix is equivalent to the value of input argument reptime.

Examples

MAR_Variance(reptime = 2, type = "V1")

SNSeg documentation built on June 22, 2024, 10:50 a.m.