sm: Scale Model

Description

This method produces a model for scale of distribution for the provided pre-estimated model. The model can be estimated either via lm or alm.

Usage

sm(object, ...)
# S3 method for default
sm(object, formula = NULL, data = NULL,
  parameters = NULL, ...)
# S3 method for lm
sm(object, formula = NULL, data = NULL, parameters = NULL,
  ...)
# S3 method for alm
sm(object, formula = NULL, data = NULL, parameters = NULL,
  ...)

Arguments

object: The pre-estimated alm or lm model.
...: Other parameters to pass to the method, including those explained in alm (e.g. parameters for optimiser).
formula: The formula for scale. It should start with ~ and contain all variables that should impact the scale.
data: The data, on which the scale model needs to be estimated. If not provided, then the one used in the object is used.
parameters: The parameters to use in the model. Only needed if you know the parameters in advance or want to test yours.

Author

Ivan Svetunkov, ivan@svetunkov.com

Details

This function is useful, when you suspect a heteroscedasticity in your model and want to fit a model for the scale of the pre-specified distribution. This function is complementary for lm or alm.

Examples

Run this code


xreg <- cbind(rnorm(100,10,3),rnorm(100,50,5))
xreg <- cbind(100+0.5*xreg[,1]-0.75*xreg[,2]+sqrt(exp(0.8+0.2*xreg[,1]))*rnorm(100,0,1),
              xreg,rnorm(100,300,10))
colnames(xreg) <- c("y","x1","x2","Noise")

# Estimate the location model
ourModel <- alm(y~.,xreg)
# Estimate the scale model
ourScale <- sm(ourModel,formula=~x1+x2)
# Summary of the scale model
summary(ourScale)