artransform: Mapping real valued parameters to stationary region

In KFAS: Kalman Filter and Smoother for Exponential Family State Space Models

Mapping real valued parameters to stationary region

Description

Function artransform transforms p real valued parameters to stationary region of pth order autoregressive process using parametrization suggested by Jones (1980). Fortran code is a converted from stats package's C-function partrans.

Usage

artransform(param)

Arguments

param

Real valued parameters for the transformation.

Value

transformed The parameters satisfying the stationary constrains.

Note

This should in theory always work, but in practice the initial transformation by tanh can produce values numerically identical to 1, leading to AR coefficients which do not satisfy the stationarity constraints. See example in logLik.SSModel on how to scope with those issues.

References

Jones, R. H (1980). Maximum likelihood fitting of ARMA models to time series with missing observations, Technometrics Vol 22. p. 389–395.

Examples

artransform(1:3)

KFAS documentation built on Sept. 8, 2023, 5:56 p.m.