best_glp_models: Best AIC/BIC Models for Specified GLP

In artfima: ARTFIMA Model Estimation

Description

This function is used by bestModels

Usage

best_glp_models(z, glp = c("ARTFIMA", "ARFIMA", "ARIMA"), p = 2, q = 2, 
   likAlg = c("exact", "Whittle"), d=0, ...)

Arguments

z	time series
glp	glp is equal to one of the following choices: "ARTFIMA", "ARFIMA" or "ARIMA"
p	maximum order of AR component
q	maximum order of MA component
likAlg	likAlg = c("exact", "Whittle")) either "exact" or "Whittle"
d	regular integer differencing parameter
...	optional arguments for artfima such as lambdaMax

Value

A list with 4 entries:

LL	log-likelihood of models
artfima_time	total time
aic	list with best aic models
bic	list with best bic models

Each of the components aic and bic is a list with three components:

bestaic	best aic models
bestbicModel	best model
aic	plausability

Similarly for the bic component.

Author(s)

A. I. McLeod

See Also

bestModels

Examples

## Not run: 
#takes about 4 minutes. Checking result for bestmodels()
z<-tseg(1000, "BJARMA11")
ansARIMA <- best_glp_models(z, glp = "ARIMA", p=2, q=2)
ansARFIMA <- best_glp_models(z, glp = "ARFIMA", p=2, q=2)
ansARTFIMA <- best_glp_models(z, glp = "ARTFIMA", p=2, q=2)
ansARIMA$bic$bic
ansARFIMA$bic$bic
ansARTFIMA$bic$bic
bestModels(z)

## End(Not run)

artfima documentation built on May 2, 2019, 1:27 p.m.