RLDM: Rational Linear Dynamic Models

This package provides tools for stationary processes with rational spectral density. It implements VARMA and state space models with methods for estimation, simulation, prediction, and model comparison.

Details

The package uses Rcpp/RcppArmadillo for performance-critical computations including Kalman filtering and recursive least squares. It depends on the sister package rationalmatrices for rational matrix operations.

Package Organization

R source files use a numeric prefix system organized by purpose/workflow:

• 01_: Model representations and classes (armamod, stspmod, rmfdmod)

• 02_: Parameter templates (tmpl_* functions, fill_template, extract_theta)

• 03_: Derived properties (autocovariance, frequency response, impulse response, spectral density, poles)

• 04_: Time series operations (solve_de, sim, prediction/forecasting)

• 05_: Estimation methods (AR, ARMA, subspace, likelihood, recursive least squares)

• 06_: Visualization (plot methods for properties and predictions)

• 07_: Model comparison metrics and diagnostics

• 08_: Utilities (print, str, data documentation, package metadata)

Getting Started

See vignettes for different user levels:

• vignette("0_getting_started") for beginner-friendly introduction

• vignette("1_case_study") for practical end-to-end workflow

• vignette("2_technical_reference") for technical details and method selection

Citation

When using RLDM in publications, please cite the relevant theory papers referenced in the vignettes. Use citation("RLDM") for package citation.

See also

Useful links:

• https://bfunovits.github.io/RLDM

• https://github.com/bfunovits/RLDM

• Report bugs at https://github.com/bfunovits/RLDM/issues

Author

Wolfgang Scherrer, Bernd Funovits Maintainer: bernd.funovits@gmail.com