‘pcts’ is an R package for modelling periodically correlated and periodically integrated time series.
Install the latest stable version of pcts from CRAN:
install.packages("pcts")You can install the development version of pcts from Github:
library(devtools)
install_github("GeoBosh/pcts")Periodic time series can be created with pcts(). Models are fitted with fitPM() and several other functions. To obtain periodic properties, such as sample periodic autocorrelations of periodic time series or theoretical periodic autocorrelations of periodic models, just call the respective functions (here autocorrelations() and partialAutocorrelations()) and they will compute the relevant property depending on the class of the argument, see the examples in the documentation.
A good place to start is the help topic pcts-package. The data from Franses (1996) are available for examples and experiments as dataFranses1996, see ?dataFranses1996. dataFranses1996 is of class "mts" (the standard R class for multivariate time series).
This is a major update of (unreleased) previous versions of ‘pcts’ which have been used in the last fifteen years or so by me and collaborators. I renamed a number of functions and other objects to improve consistency and to follow the current convention not to use dot (.) in function names. Also, I have temporarily removed some functions which will be put back after suitable revision.