We provide a toolbox to fit a continuous-time fractionally integrated ARMA process (CARFIMA) on univariate and irregularly spaced time series data via frequentist or Bayesian machinery. A general-order CARFIMA(p, H, q) model for p>q is specified in Tsai and Chan (2005)<doi:10.1111/j.1467-9868.2005.00522.x> and it involves p+q+2 unknown model parameters, i.e., p AR parameters, q MA parameters, Hurst parameter H, and process uncertainty (standard deviation) sigma. The package produces their maximum likelihood estimates and asymptotic uncertainties using a global optimizer called the differential evolution algorithm. It also produces their posterior distributions via Metropolis within a Gibbs sampler equipped with adaptive Markov chain Monte Carlo for posterior sampling. These fitting procedures, however, may produce numerical errors if p>2. The toolbox also contains a function to simulate discrete time series data from CARFIMA(p, H, q) process given the model parameters and observation times.
| Version: | 1.0.0 |
| Depends: | R (≥ 2.2.0) |
| Imports: | MASS (≥ 7.3-47), DEoptim (≥ 2.2-4), numDeriv (≥ 2016.8-1), truncnorm (≥ 1.0-7), invgamma (≥ 1.1) |
| Published: | 2017-10-23 |
| Author: | Hyungsuk Tak and Henghsiu Tsai |
| Maintainer: | Hyungsuk Tak <hyungsuk.tak at gmail.com> |
| License: | GPL-2 |
| NeedsCompilation: | no |
| In views: | TimeSeries |
| CRAN checks: | carfima results |
| Reference manual: | carfima.pdf |
| Package source: | carfima_1.0.0.tar.gz |
| Windows binaries: | r-devel: carfima_1.0.0.zip, r-release: carfima_1.0.0.zip, r-oldrel: carfima_1.0.0.zip |
| OS X binaries: | r-release: carfima_1.0.0.tgz, r-oldrel: carfima_1.0.0.tgz |
Please use the canonical form https://CRAN.R-project.org/package=carfima to link to this page.