caschrono (version 2.4)

armaselect: Minic method

Description

armaselect implements the MINIC (Minimum Information Criterion) identification method and returns the nbmod best ARMA models, with respect to the Schwarz's Bayesian Criterion (sbc).

Usage

armaselect(y, max.p = 15, max.q = 15, nbmod = 10)

Value

A matrix with nbmod rows and 3 columns (values of p, q and sbc)

Arguments

y

a time series

max.p

an integer, the maximum value for the autoregressive component, p

max.q

an integer, the maximum value for the moving average component, q

nbmod

an integer, the number of models that will be returned (nbmod may be lower than max.p x max.q).

Author

Yves Aragon

Examples

set.seed(4123)
n2 <- 210
yc <- arima.sim(n = 200, list(ar = -0.8, ma = c(-0.3, 0.6)),
 sd = sqrt(1.5))
yc <- yc - 10
armaselect(yc, nbmod = 5)