Skip to contents

Compute standard errors

se.Rd

Compute standard errors.

Usage

se(object, ...)

# S4 method for SampleAutocorrelations
vcov(object, assuming = "iid", maxlag = maxLag(object), ...)

Arguments

object

an object containing estimates, such as a fitted model.

...

further arguments for vcov.

assuming

under what assumptions to do the computations? Currently can be "iid", "garch", a fitted model, or a theoretical model, see Details.

maxlag

maximal lag to include

Details

se is a convenience function for the typical case where only the square root of the diagonal of the variance-covariance matrix is needed.

The method for vcov gives the variance-covariance matrix of the first maxlag autocorrelation coefficients in the object. The result depends on the underlying assumptions and the method of calculation. These can be specifyed with the additional arguments.

Argument "assuming" can be though also as specifying a null hypothesis. Setting it to "iid" or "garch" corresponds to strong white noise (iid) and weak white noise, respectively.

Setting "assuming" to an ARMA model (theoretical or fitted) specifies that as the null model.

Note: The method for vcov is not finalised yet. It is used by a method for confint. Bug reports and requests on the github repo may bring this closer to the top of my task list.

Value

for se, a numeric vector giving standard errors;

for the vcov method, a square matrix

See also

link{confint}, vcov