bin_by_iq_def: Iterative Quantile Binning New Data from defined bins

In kmaurer/binsemble:

Description

New observations selected from the same population as the data used to build bin definitions may fall just outside the bins. If we wish to include nearby values we can either allow outer bins to be extended (this function) or to leave the outer bins unbounded.

Usage

bin_by_iq_def(bin_def, new_data, output = "data", strict = FALSE)

Arguments

new_data	Data frame with column names matching the binned columns from bin-training data
output	Matches format of iterative_quant_bin and inherets properties from iqnn if applicable "data","both"
strict	TRUE/FALSE: If TRUE Observations must fall within existing bins to be assigned; if FALSE the outer bins in each dimension are unbounded to allow outlying values to be assigned.
iq_def	Iterative quantile binning definition list

Value

updated binning definition with bins extended by tolerance values

Examples

withhold_index <- c(1,2,51,52,101,102)
iq_def <- iterative_quant_bin(data=iris[-withhold_index,], bin_cols=c("Sepal.Length","Sepal.Width","Petal.Width"),
                              nbins=c(3,2,2), output="definition")
bin_by_iq_def(bin_def=iq_def, new_data=iris[withhold_index,], output="data")

kmaurer/binsemble documentation built on May 7, 2019, 9:50 p.m.