The Fisher large-sample information matrix per observation for
the p coefficients in an AR(p) is computed.
Usage
InformationMatrixAR(phi)
Arguments
phi
vector of length p corresponding to the AR(p) coefficients
Value
a p-by-p Toeplitz matrix, p = length(phi)
Details
The Fisher information matrix is computed as the covariance matrix
of an AR(p) process with coefficients given in the argument
phi and with unit innovation variance.
The TacvfAR function is used to compute the necessary
autocovariances.
FitAR uses InformationMatrixAR to obtain estimates
of the standard errors for the estimated parameters in the case
of the full AR(p) model.
References
McLeod, A.I. and Zhang, Y. (2006).
Partial autocorrelation parameterization for subset autoregression.
Journal of Time Series Analysis, 27, 599-612.