BWselection: BackWard Selection

In xuetao666/BWselection: Backward Selection

Description

Perform Backward selection on regression models

Usage

BWselection(
  data,
  qvarlist = NULL,
  lvarlist = NULL,
  spvarlist = NULL,
  spclist = NULL,
  catvarlist = NULL,
  outcome,
  type = "lm",
  sig = 0.05,
  complete_case = TRUE,
  track = "Simple"
)

Arguments

data	input Data used for the regression
qvarlist	Variable list in vector format, including all the variables needed for the quadratic form
lvarlist	Variable list in vector format, including all the variables in the continuous linear form
spvarlist	variable list in vector format, including all the variables for the spline form
spclist	if spvarlist is not empty, specify the outpoints for the spline term
catvarlist	variable list in the vector format, including all the categorical variables
outcome	Outcome variable for the regression model
type	type of regression model, default is lm, logistic regression use type="logistic"
sig	significant level used for exit the model, default is 0.05
complete_case	whether to use the complete case for all the variables in the model, default is TRUE
track	Whether print the selection process, could select from "Simple","All", "None"

Value

a list of original formula, final fit model and final fit formula

Examples

#Load NHANES DATA
library(NHANES)
data("NHANES")
data=NHANES

#Created variable lists for the model selection
qvarlist<-c("BPSysAve","SleepHrsNight","TotChol")
lvarlist<-c("Age","BPDiaAve","Weight","Height")
spvarlist<-c("Pulse","DirectChol")
spclist<-c(70,1.2)
catvarlist<-c("Depressed","Marijuana","Gender")

#Run the model selection, this will return a list
fit_test1=BWselection(data=NHANES,qvarlist=qvarlist,lvarlist=lvarlist,spvarlist=spvarlist,
                     spclist=spclist,catvarlist=catvarlist,outcome="BMI",type="lm",sig=0.05,complete_case=TRUE)

xuetao666/BWselection documentation built on Dec. 23, 2021, 7:13 p.m.