cat_new_class: Clubbing class of categorical variables with low population...

In scorecardModelUtils: Credit Scorecard Modelling Utils

Description

The function groups classes of categorical variables, which have population percentage less than a threshold, with another class of similar event rate. If a class of exactly same event rate is not available, it is clubbed with the one having a higher event rate closest to it.

Usage

cat_new_class(base, target, cat_var_name, threshold, event = 1)

Arguments

base	input dataframe
target	column / field name for the target variable to be passed as string (must be 0/1 type)
cat_var_name	column name or array of column names of categorical variable on which the operation is to be done, to be passed as string
threshold	threshold population percentage below which the class will be considered to be be clubbed with another class, to be provided as decimal/fraction
event	(optional) the event class, to be passed as 0 or 1 (default is 1)

Value

The function returns an object of class "cat_new_class" which is a list containing the following components:

base_new	a dataframe after clubbing low percentage classes with another class of similar or closest but higher event rate
cat_class_new	a dataframe with mapping between original classes and new clubbed classes (if any)

Author(s)

Arya Poddar <[e-mail]>

Kanishk Dogar <[e-mail]>

Examples

data <- iris[1:110,]
data$Species <- as.character(data$Species)
data$Y <- sample(0:1,size=nrow(data),replace=TRUE)
data_newclass <- cat_new_class(base = data,target = "Y",cat_var_name = "Species",threshold = 0.1)

Example output

scorecardModelUtils documentation built on May 2, 2019, 9:59 a.m.