optimal_bin: Optimally discretize variables in your training and test...

In awstringer/modellingTools: A Collection of Useful Custom Tools for D&A and Modelling

Description

This function will perform supervised discritization of all variables in the supplied training, and optionally test, datasets. This utilizes the binning functionality of modellingTools::simple_bin combined with the optimal bin calculations performed in smbinning::smbinning. Note that no filtering is done on the resulting binning structure; there may be pure bins, non-monotonic Weights of Evidience, etc. This is left to the user- the package provides a tool-set for dealing with any such concerns.

Usage

optimal_bin(train, response, exclude_vars = NULL, include_vars = NULL)

Arguments

train	training set
response	a string naming the response variable; must be 0/1 and coercible to factor
exclude_vars	variables to exclude (e.g. the target, or the row ID)
include_vars	if you only want certain variables binned, you may specify them directly instead of excluding all other variables

Value

a list containing the following elements: iv: a dataframe containing the variables and their information values, sorted in descending order train: a tbl_df containing the same variables as train, with the appropriate ones binned (per exclude_vars or include_vars) test: if test is NULL, then NULL; else a tbl_df containing the same variables as test, binned in the same manner as train.

awstringer/modellingTools documentation built on May 11, 2019, 4:11 p.m.