lsar.chgpt {TSSS}

Estimation of the Change Point

Description

Precisely estimate a change point of subinterval for locally stationary AR model.

Usage

lsar.chgpt(y, max.arorder = 20, subinterval, candidate, plot = TRUE, ...)

Arguments

y

a univariate time series.

max.arorder

highest order of AR model.

subinterval

a vector of the form c(n0, ne) which gives a start and end point of time interval used for model fitting.

candidate

a vector of the form c(n1, n2) which gives minimum and maximum of the candidate for change point.

n0+2k < n1 < n2+k < ne, ( k is max.arorder )

plot

logical. If TRUE (default), y[n0:ne] and aic are plotted.

...

graphical arguments passed to the plot method.

Value

An object of class "chgpt" which has a plot method. This is a list with the following components:

aic

AICs of the AR models fitted on [n1, n2].

aicmin

minimum AIC.

change.point

estimated change point.

subint

information about the original sub-interval.

References

Kitagawa, G. (2020) Introduction to Time Series Modeling with Applications in R. Chapman & Hall/CRC.

Examples

# seismic data
data(MYE1F)
lsar.chgpt(MYE1F, max.arorder = 10, subinterval = c(200, 1000),
           candidate = c(400, 800))

lsar.chgpt(MYE1F, max.arorder = 10, subinterval = c(600, 1400),
           candidate = c(800, 1200))