make_numeric: Turning a non-numeric variable into a numeric one

In CARRoT: Predicting Categorical and Continuous Outcomes Using One in Ten Rule

Turning a non-numeric variable into a numeric one

Description

Function which turns a single categorical (non-numeric) variable into a numeric one (or several) by introducing dummy '0'/'1' variables.

Usage

make_numeric(vari, outcome, ra,mode)

Arguments

vari	array of values to be transformed
outcome	TRUE/FALSE indicates whether the variable vari is an outcome (TRUE) or a predictor (FALSE)
ra	indices of the input array vari which indicate which values will be transformed
mode	'binary' (logistic regression), 'multin' (multinomial regression)

Details

This function is essentially a standard way to turn categorical non-numeric variables into numeric ones in order to run a regression

Value

Returned value is an M x N matrix where M is the length of the input array of indices ra and N is length(vari)-1.

Examples

#creating a non-numeric set

a<-t(rmultinom(100,1,c(0.2,0.3,0.5)))%*%c(1,2,3)

a[a==1]='red'
a[a==2]='green'
a[a==3]='blue'

#running the function

make_numeric(a,FALSE,sample(1:100,50),"linear")

make_numeric(a,TRUE,sample(1:100,50))

CARRoT documentation built on Oct. 14, 2023, 1:06 a.m.