The function provides the applicability of different types of transformation techniques for the data to which the Ata method will be applied.
The ATA.Transform
function works with many different types of inputs.
ATA.Transform(
X,
tMethod = c("Box_Cox", "Sqrt", "Reciprocal", "Log", "NegLog", "Modulus",
"BickelDoksum", "Manly", "Dual", "YeoJohnson", "GPower", "GLog"),
tLambda,
tShift = 0,
bcMethod = c("loglik", "guerrero"),
bcLower = 0,
bcUpper = 5
)
A list object consists of transformation parameters and transformed data.
ATA.Transform
is a list containing at least the following elements:
trfmX : Transformed data
tLambda : Box-Cox power transformation family parameter
tShift : Box-Cox power transformation family shifting parameter
a numeric vector or time series of class ts
or msts
for in-sample.
Box-Cox power transformation family is consist of "Box_Cox", "Sqrt", "Reciprocal", "Log", "NegLog",
"Modulus", "BickelDoksum", "Manly", "Dual", "YeoJohnson", "GPower", "GLog" in ATAforecasting package. If the transformation process needs shift parameter,
ATA.Transform
will calculate required shift parameter automatically.
Box-Cox power transformation family parameter. Default is NULL. When lambda is set as NULL, required "lambda" parameter will be calculated automatically based on "bcMethod, bcLower, and bcUpper".
Box-Cox power transformation family shifting parameter. Default is 0. When "transform.method" is selected, required shifting parameter will be calculated automatically according to dataset.
Choose method to be used in calculating lambda. "loglik" is default. Other method is "guerrero" (Guerrero, V.M. (1993)).
Lower limit for possible lambda values. The lower value is limited by -5. Default value is 0.
Upper limit for possible lambda values. The upper value is limited by 5. Default value is 1.
#'tukey1957ATAforecasting
#'boxcox1964ATAforecasting
#'manly1976ATAforecasting
#'johndraper1980ATAforecasting
#'bickeldoksum1982ATAforecasting
#'sakia1992ATAforecasting
#'guerrero1993ATAforecasting
#'yeojohn2000ATAforecasting
#'glog2002ATAforecasting
#'neglog2005ATAforecasting
#'yang2006ATAforecasting
#'gpower2013ATAforecasting