methods: Methods for 'mcsSubset' objects

In mcsSubset: MCS: Variable Subset Selection in Linear Regression

Description

Extractor methods for mcsSubset objects.

Usage

## S3 method for class 'mcsSubset'
variable.names(object, size = NULL, best = 1,
..., .full = FALSE, .neg = FALSE)
## S3 method for class 'mcsSubset'
formula(x, ...)
## S3 method for class 'mcsSubset'
model.frame(formula, ...)
## S3 method for class 'mcsSubset'
model.matrix(object, size = NULL, best = 1, ...)
## S3 method for class 'mcsSubset'
refit(object, ...)
## S3 method for class 'mcsSubset'
coef(object, ...)
## S3 method for class 'mcsSubset'
vcov(object, ...)
## S3 method for class 'mcsSubset'
fitted(object, ...)
## S3 method for class 'mcsSubset'
residuals(object, ...)
## S3 method for class 'mcsSubset'
deviance(object, size = NULL, best = 1, ...)
## S3 method for class 'mcsSubset'
logLik(object, size = NULL, best = 1, ..., df)
## S3 method for class 'mcsSubset'
AIC(object, size = NULL, best = 1, ..., k = NULL)
## S3 method for class 'mcsSubset'
BIC(object, size = NULL, best = 1, ...)

Arguments

object, formula, x	An object of class mcsSubset.
size	The subset size.
best	The subset rank.
df	Degrees of freedom.
k	AIC penalty.
...	Forwarded arguments.
.full	For internal use.
.neg	For internal use.

Details

The extractor methods work for mcsSubset objects that have been generated using the formula interface.

The information is extracted from the model refitted to a given size. If a method is not available, refit can always be called explicitly to obtain a new lm object of the desired size or rank. In some cases, the size and best arguments may be a vector specifying multiple values.

The method refit returns an lm object fitted to the desired size and rank.

See Also

mcsSubset, refit.

Examples

## load data
data("AirPollution", package = "mcsSubset")

## fit subsets (5 best subsets per size)
xs <- mcsSubset(mortality ~ ., data = AirPollution, nbest = 5)

## extract information (for subset of size 3)
coef(xs, size = 3)
vcov(xs, size = 3)
residuals(xs, size = 3)
fitted(xs, size = 3)
model.matrix(xs, size = 3)

## summarize (BIC)
sx <- summary(xs, penalty = log(nrow(AirPollution)))

## extract information (for best BIC subset)
deviance(sx)
logLik(sx)
AIC(sx)

## refit model
lm5 <- refit(sx, size=5)
summary(lm5)
## (Note that the p-values are not valid due to model selection.)

## select 5 best subsets using AIC
xs <- mcsSubset(mortality ~ ., data = AirPollution, penalty = 2, nbest = 5)

## summarize
summary(xs)

## extract deviance for best subset
deviance(xs)

## extract BIC for all subsets
BIC(xs, best = 1:5)

mcsSubset documentation built on May 2, 2019, 6:50 p.m.