FPE.varlse {bvhar}R Documentation

Final Prediction Error Criterion of Multivariate Time Series Model

Description

Compute FPE of VAR(p), VHAR, BVAR(p), and BVHAR

Usage

## S3 method for class 'varlse'
FPE(object, ...)

## S3 method for class 'vharlse'
FPE(object, ...)

Arguments

object

Model fit

...

not used

Details

Let Σ~e\tilde{\Sigma}_e be the MLE and let Σ^e\hat{\Sigma}_e be the unbiased estimator (covmat) for Σe\Sigma_e. Note that

Σ~e=sknΣ^e\tilde{\Sigma}_e = \frac{s - k}{n} \hat{\Sigma}_e

Then

FPE(p)=(s+ksk)mdetΣ~eFPE(p) = (\frac{s + k}{s - k})^m \det \tilde{\Sigma}_e

Value

FPE value.

References

Lütkepohl, H. (2007). New Introduction to Multiple Time Series Analysis. Springer Publishing.


[Package bvhar version 2.0.1 Index]