ml_mean_target_encoder: Mean Target encoding is the process of replacing a...

In ddzr/XL.RPackage: XL R Package

Description

This function convert input columns to mean target encoding columns, inspired from https://www.kaggle.com/ogrellier/python-target-encoding-for-categorical-features

Usage

ml_mean_target_encoder(
  mean_target_colnames,
  target_colname,
  training_data,
  test_data = NULL,
  fold = NULL,
  min_samples_leaf = 0,
  smoothing = 1,
  impute_missing = T
)

Arguments

mean_target_colnames	list of column names that needs mean target encoding
target_colname	target column name, should be numeric column
training_data	dataset we used to build mean targeting table
test_data	dataset we used to apply mean target encoding
fold	list of fold index for training dataset, suggest to use caret::createFolds
min_samples_leaf	minimum quantity per column value we consider needs to be calculate mean.
smoothing	used to control rate of transition between the particular column level’s posterior probability and the grand target mean, default is set to 1 to ignore this effect
impute_missing	determine if we want to impute missing when new dataste levels could not be found in mapping table, defualt is to impute

Value

a list that contains datasets that haven been applied mean target encoding and also a mapping table named calculate_columns_mean_table to use for ml_mean_target_apply function ta apply mean targeting encoding on new dataset

See Also

ml+mean_target_apply

ddzr/XL.RPackage documentation built on June 8, 2020, 12:36 a.m.