References & Citations
Statistics > Computation
Title: The gmwm R package: a comprehensive tool for time series analysis from state-space models to robustness
(Submitted on 15 Jul 2016)
Abstract: The gmwm R package for inference on time series models is mainly based on the quantity called wavelet variance which is derived from a wavelet decomposition of a time series. This quantity provides a means to summarize and graphically represent the features of time series in order to identify possible models. Moreover, it is used as a moment condition for model estimation through the generalized method of wavelet moments. Based on the latter method, this package not only provides an alternative method to estimate classical ARMA models but also delivers a general framework for the robust estimation of many time series models as well as a quick and efficient estimation of many linear state-space models.
Submission history
From: Roberto Molinari Mr [view email][v1] Fri, 15 Jul 2016 15:03:42 GMT (908kb,D)
Link back to: arXiv, form interface, contact.