init-vars: Initial Parameter Values
In stsm: Structural Time Series Models

Description Usage Arguments Details Value References See Also

This function computes initial variance parameters to be used as starting parameter values in an optimization procedure.

1	init.vars(model, debug = FALSE)

`model`	an object of class `stsm`.
`debug`	logical. If `TRUE`, the correctness if the result is double-checked.

As mentioned in Harvey (1989), the frequency domain representation of the structural model suggests using a linear regression to compute initial variance parameters from which to start an optimization procedure. The variable 2π times the periodogram is regressed on the constant terms of the spectral generating function of the model.

A list containing the initial variance parameters and the output of the linear regression.

Harvey, A. C. (1989). Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge University Press.

maxlik.fd, stsm.sgf.

stsm documentation built on May 2, 2019, 7:39 a.m.