baggedClusterETS: Perform BaggedClusterETS method

In tiagomendesdantas/tshacks: This packages brings to you time series hacks to enhance foreacast and other things

Description

BaggedClusterETS method

Usage

baggedClusterETS(y, cores = detectCores() - 1, nclusters = 5,
  distance = "EUCL", silhouette = F, h_pseudo = NULL,
  boot_samples = 1000, ...)

Arguments

y	a time series
cores	(optional) number of cores used to do the calculations
nclusters	(optional) number of clusters used. If not selected it uses the Silhouette method to define it
distance	distance function used to create the clusters. Default is the Euclidian distance but one can choose among many from the TSclust package
silhouette	If TRUE the method performs the Silhouette method and ignores the nclusters selection
h_pseudo	number of observations to the validation set that will be used to select the best series
boot_samples	number of bootstrap series generated to perform the cluster phase
...	other parameters that affect the ETS estimation

Value

a list with several elements such as the generated bootstrapped series, fitted method to each of the bootstrapped series and so on

Examples

## Not run: 
model <- baggedCLusterETS(gas)
plot(forecast(model))

## End(Not run)

tiagomendesdantas/tshacks documentation built on Sept. 2, 2020, 1:08 a.m.