Description
Tools for the Analysis of Weak ARMA Models
Description
Numerous time series admit autoregressive moving average (ARMA) representations, in which the errors are uncorrelated but not necessarily independent. These models are called weak ARMA by opposition to the standard ARMA models, also called strong ARMA models, in which the error terms are supposed to be independent and identically distributed (iid). This package allows the study of nonlinear time series models through weak ARMA representations. It determines identification, estimation and validation for ARMA models and for AR and MA models in particular. Functions can also be used in the strong case. This package also works on white noises by omitting arguments 'p', 'q', 'ar' and 'ma'. See Francq, C. and Zakoïan, J. (1998) <doi:10.1016/S0378-3758(97)00139-0> and Boubacar Maïnassara, Y. and Saussereau, B. (2018) <doi:10.1080/01621459.2017.1380030> for more details.
README.md
weakARMA
The goal of weakARMA is to allows the study of nonlinear time series models through weak ARMA representations.
Installation (Gitlab)
Current released
You can install the released version of weakARMA from PLMlab with:
install.packages("remotes")
remotes::install_gitlab("jrolland/weakARMA", host="https://plmlab.math.cnrs.fr")
Development version
You can install the currently developed version of weakARMA from PLMlab with:
install.packages("remotes")
remotes::install_git("https://plmlab.math.cnrs.fr/jrolland/weakARMA.git", ref="develop")
Installation (CRAN)
CRAN package is available. You can install the released version of weakARMA from CRAN with:
install.packages("weakARMA")
Example
This is a basic example which shows you how to solve a common problem:
library(weakARMA)
## basic example code