GEVStableGarch (version 1.1)

gsSelect:

Selects the best model according to goodness-of-fit criteria

Description

This function estimates ARMA-GARCH/APARCH models with varying order and returns the one with the smallest goodness-of-fit criteria

Usage

gsSelect(data, order.max = c(1, 1, 1, 1), 
selection.criteria = c("AIC", "AICc", "BIC"), is.aparch = FALSE, 
cond.dist = c("stableS0", "stableS1", "stableS2", "gev", 
"gat", "norm", "std", "sstd", "skstd", "ged"), 
include.mean = TRUE, 
algorithm = c("sqp", "sqp.restriction", "nlminb", "nlminb+nm"), ...)

Arguments

data

Data with the time series to be estimated. It must be a numeric vector not contain NA, NULL or Inf values.

order.max

Maximum order of models to search. It must by a vector of the type c(mMax, nMax, pMax, qMax) with each entry representing the maximum order of the model that will be fitted when searching for the best model.

selection.criteria

The goodness-of-fit criterion to be used when searching for the best model. Three different criterias are allowed: AIC, AICc or BIC. For more information see Brockwell and Davis (1996).

is.aparch

Boolean variable indicating whether to search for ARMA-GARCH or ARMA-APARCH models.

cond.dist

a character string naming conditional distribution of innovations. The package was created to accept the following distributions: "stableS0", "stableS1", "stableS2", "gev" and "GAt". Other common distributions are also possible such as the "norm", "std", "sstd", "skstd", "ged".

include.mean

This is a boolean variable. It intercept is TRUE than we estimate the model with intercept.

algorithm

The algorithm to be used to search for the optimum value. The current version of the GEVStableGarch package implements four different optimization procedures, namely the "sqp", "sqp.restriction" (enforce stationarity) , "nlminb" and "nlminb+nm". See the details section for more information about the estimation algorithms.

...

Additional parameters to be passed to function gsFit during estimation.

Value

Returns a S4 object of class "GEVSTABLEGARCH" with the best model. See GEVSTABLEGARCH-class for details.

References

Brockwell, P.J., Davis, R.A. (1996). Introduction to Time Series and Forecasting. Springer, New York.

Examples

# Best ARMA-GARCH model within the range ARMA(0,0)-GARCH(1,0) to ARMA(0,0)-GARCH(1,1)
# using the Corrected Akaike Information Criteria (AICc)
data(dem2gbp)
x = dem2gbp[,1]
model = gsSelect (data = x, order.max = c(0,0,1,1), is.aparch = FALSE, 
          algorithm = "sqp", cond.dist = "gev", selection.criteria = "AIC", 
          include.mean = FALSE)