rareEncode: rareEncode

In AnotherSamWilson/helperFuncs: Helper Functions for common manipuations

Description

Groups rare levels together, where 'rare' is defined by either minPerc or minCount.

Usage

rareEncode(dt, vars, minPerc = NULL, minCount = NULL,
  newString = "rareGroup", encodeNA = FALSE, allowNewLevels = FALSE)

Arguments

dt	Dataset to create object on.
vars	variables you want to include in the encoding.
minPerc	Minimum percentage of the total population required for a group to be considered 'rare'. For example, if minPerc = 0.1, and Gender = 'Female' only makes up 5 Female would be replaced with the string in newString.
minCount	Can be provided instead of minPerc if you have a better idea of credibility in terms of total samples required. Levels with counts below this value will be grouped into newstring.
newString	Value you want to group rare levels together as.
encodeNA	If FALSE, NAs will remain NA even if they would have been grouped. If TRUE, NAs will be grouped, but only if they are 'rare'. If they are not rare, NAs will be untouched.
allowNewLevels	Similar to frequencyEncode parameter. This groups all new levels in with newString.

Value

Rare Encoded Object

AnotherSamWilson/helperFuncs documentation built on Oct. 1, 2019, 8:51 p.m.