stoch.reg: Frequency Domain Stochastic Regression

View source: R/stoch.reg.R

stoch.regR Documentation

Frequency Domain Stochastic Regression

Description

Performs frequency domain stochastic regression discussed in Chapter 7.

Usage

stoch.reg(xdata, cols.full, cols.red=NULL, alpha, L, M, plot.which, col.resp=NULL, ...)

Arguments

xdata

data matrix with the last column being the response variable

cols.full

specify columns of data matrix that are in the full model

cols.red

specify columns of data matrix that are in the reduced model (use NULL if there are no inputs in the reduced model)

alpha

test size; number between 0 and 1

L

odd integer specifying degree of smoothing

M

number (integer) of points in the discretization of the integral

plot.which

coh or F.stat, to plot either the squared-coherencies or the F-statistics, respectively

col.resp

specify column of the response variable if it is not the last column of the data matrix

...

additional graphic arguments

Details

This function computes the spectral matrix, F statistics and coherences, and plots them. Returned as well are the coefficients in the impulse response function.

Enter, as the argument to this function, the full data matrix, and then the labels of the columns of input series in the "full" and "reduced" regression models - enter NULL if there are no inputs under the reduced model.

If the response variable is the LAST column of the data matrix, it need not be specified. Otherwise specify which column holds the responses as col.resp.

Other inputs are alpha (test size), L (smoothing), M (number of points in the discretization of the integral) and plot.which = "coh" or "F", to plot either the coherences or the F statistics.

Value

power.full

spectrum under the full model

power.red

spectrum under the reduced model

Betahat

regression parameter estimates

eF

pointwise (by frequency) F-tests

coh

coherency

Note

See Example 7.1 of the text. The script is based on code that was contributed by Professor Doug Wiens, Department of Mathematical and Statistical Sciences, University of Alberta.

Author(s)

D.S. Stoffer

References

You can find demonstrations of astsa capabilities at FUN WITH ASTSA.

The most recent version of the package can be found at https://github.com/nickpoison/astsa/.

In addition, the News and ChangeLog files are at https://github.com/nickpoison/astsa/blob/master/NEWS.md.

The webpages for the texts and some help on using R for time series analysis can be found at https://nickpoison.github.io/.


astsa documentation built on May 29, 2024, 10:29 a.m.