data_screen: Data screening

In MCKnaus/causalDML: Causal Double Machine Learning

Data screening

Description

This function takes a matrix of data and removes 1. Variables without variation 2. Dummy variables where one group is nearly empty (optional in one of both treatment groups) 3. Redundant (highly correlated variables)

Usage

data_screen(data, treat = NULL, bin_cut = 0.01, corr_cut = 0.99, print = FALSE)

Arguments

data	Matrix the variables to be screened.
treat	Optional binary treatment vector if screening should be done within treatment groups
bin_cut	Cut-off fraction under which nearly empty binary variables should be removed. Default 0.01.
corr_cut	Cut-off above which highly correlated varialbes should be removed. Default 0.99.
print	Shows details about the reomved variables at each step

Value

Screened matrix

MCKnaus/causalDML documentation built on Aug. 19, 2023, 5:47 p.m.