boot.select.gen: Perform a Step of FSS

In nkissel/ModelPath: Model Path Selection

Description

Internal function that performs 1 step of forward stability selection. Returns a table of selection counts.

Usage

boot.select.gen(
  resp.name,
  c.vars,
  myframe,
  type,
  r,
  f,
  model,
  fun.args,
  pred.args
)

Arguments

resp.name	matrix of MPS data
c.vars	matrix of MPS data
myframe	data.frame holding explanatory and response data
type	'reg' or 'class' for regression and classification. Not needed if 'f' is specified.
r	maximum cell count
f	loss function
model	name of modeling method; must be name of the function called in R
fun.args	list of arguments that are passed to modeling function
pred.args	list of arguments that are passed to model predict function

Examples

library(MASS)
n<-100
p<-10
x<-mvrnorm(n,rep(0,p),diag(p))
signal<-rowSums(x[,1:3])
noise<-rnorm(n,0,1/4)
y<-signal+noise
mydata<-data.frame(x,y)
boot.select.gen(resp.name='y',c.vars=NULL,myframe=mydata,B=100,model='lm')

nkissel/ModelPath documentation built on Feb. 13, 2021, 12:53 a.m.