MDLP discretization in python

mdlp.R

# A benchmark I'm using to show that my results from MDLP are correct
# Using the R package "discretization" for comparison
library(discretization)
data(iris)
mdlp(iris)$Disc.data

mdlp_sample.py

#!/usr/bin/env python
"""NOTE: sklearn.preprocessing does not have MDLP yet. It is a feature
that I am trying to add in my pull request.
"""

from sklearn.datasets import load_iris
from sklearn.preprocessing import MDLP

iris = load_iris()
X = iris.data
Y = iris.target

"""My version of MDLP allows users to also specify which columns to
discretize. The R function assumes all columns are continuous.
I don't use this feature here, because the iris dataset contains only
continuous attributes.
"""
mdlp = MDLP()
conv_X = mdlp.fit_transform(X, Y)

"""Unlike the mdlp function in R, I allow users to specify the
minimum depth they would like to discretize, invalidating the
minimum description length principal. This could be useful sometimes.

The minimum depth will sometimes not be achieved, especially if the
algorithm can find a cut with zero entropy.
"""
mdlp = MDLP(min_depth=100)
conv_X = mdlp.fit_transform(X, Y)

python_output.txt

[[1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 1 1 1]
 [2 2 1 1]
 [2 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [2 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 1 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [1 2 1 1]
 [2 2 2 2]
 [2 2 2 2]
 [2 2 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 2 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [1 1 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 2 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 2 2 3]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 3 2]
 [2 1 2 2]
 [2 1 2 2]
 [1 1 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 1 3 2]
 [1 1 2 2]
 [2 2 2 2]
 [2 2 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [1 1 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [2 1 2 2]
 [1 1 2 2]
 [2 1 2 2]
 [2 2 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [1 1 2 2]
 [2 1 3 3]
 [2 1 3 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 1 3 2]
 [2 2 3 3]
 [2 1 2 3]
 [2 1 3 3]
 [2 1 2 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 1 2 3]
 [2 1 2 3]
 [2 1 3 3]
 [2 1 3 2]
 [2 1 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 1 3 2]
 [2 1 3 2]
 [2 1 3 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 1 2 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 2 3 3]
 [2 2 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 1 3 3]
 [2 2 3 3]
 [2 1 3 3]]

R_output.csv

"Sepal.Length" "Sepal.Width" "Petal.Length" "Petal.Width" "Species"
"1" 1 3 1 1 "setosa"
"2" 1 2 1 1 "setosa"
"3" 1 2 1 1 "setosa"
"4" 1 2 1 1 "setosa"
"5" 1 3 1 1 "setosa"
"6" 1 3 1 1 "setosa"
"7" 1 3 1 1 "setosa"
"8" 1 3 1 1 "setosa"
"9" 1 1 1 1 "setosa"
"10" 1 2 1 1 "setosa"
"11" 1 3 1 1 "setosa"
"12" 1 3 1 1 "setosa"
"13" 1 2 1 1 "setosa"
"14" 1 2 1 1 "setosa"
"15" 2 3 1 1 "setosa"
"16" 2 3 1 1 "setosa"
"17" 1 3 1 1 "setosa"
"18" 1 3 1 1 "setosa"
"19" 2 3 1 1 "setosa"
"20" 1 3 1 1 "setosa"
"21" 1 3 1 1 "setosa"
"22" 1 3 1 1 "setosa"
"23" 1 3 1 1 "setosa"
"24" 1 2 1 1 "setosa"
"25" 1 3 1 1 "setosa"
"26" 1 2 1 1 "setosa"
"27" 1 3 1 1 "setosa"
"28" 1 3 1 1 "setosa"
"29" 1 3 1 1 "setosa"
"30" 1 2 1 1 "setosa"
"31" 1 2 1 1 "setosa"
"32" 1 3 1 1 "setosa"
"33" 1 3 1 1 "setosa"
"34" 1 3 1 1 "setosa"
"35" 1 2 1 1 "setosa"
"36" 1 2 1 1 "setosa"
"37" 1 3 1 1 "setosa"
"38" 1 3 1 1 "setosa"
"39" 1 2 1 1 "setosa"
"40" 1 3 1 1 "setosa"
"41" 1 3 1 1 "setosa"
"42" 1 1 1 1 "setosa"
"43" 1 2 1 1 "setosa"
"44" 1 3 1 1 "setosa"
"45" 1 3 1 1 "setosa"
"46" 1 2 1 1 "setosa"
"47" 1 3 1 1 "setosa"
"48" 1 2 1 1 "setosa"
"49" 1 3 1 1 "setosa"
"50" 1 2 1 1 "setosa"
"51" 3 2 2 2 "versicolor"
"52" 3 2 2 2 "versicolor"
"53" 3 2 3 2 "versicolor"
"54" 1 1 2 2 "versicolor"
"55" 3 1 2 2 "versicolor"
"56" 2 1 2 2 "versicolor"
"57" 3 2 2 2 "versicolor"
"58" 1 1 2 2 "versicolor"
"59" 3 1 2 2 "versicolor"
"60" 1 1 2 2 "versicolor"
"61" 1 1 2 2 "versicolor"
"62" 2 2 2 2 "versicolor"
"63" 2 1 2 2 "versicolor"
"64" 2 1 2 2 "versicolor"
"65" 2 1 2 2 "versicolor"
"66" 3 2 2 2 "versicolor"
"67" 2 2 2 2 "versicolor"
"68" 2 1 2 2 "versicolor"
"69" 3 1 2 2 "versicolor"
"70" 2 1 2 2 "versicolor"
"71" 2 2 3 3 "versicolor"
"72" 2 1 2 2 "versicolor"
"73" 3 1 3 2 "versicolor"
"74" 2 1 2 2 "versicolor"
"75" 3 1 2 2 "versicolor"
"76" 3 2 2 2 "versicolor"
"77" 3 1 3 2 "versicolor"
"78" 3 2 3 2 "versicolor"
"79" 2 1 2 2 "versicolor"
"80" 2 1 2 2 "versicolor"
"81" 1 1 2 2 "versicolor"
"82" 1 1 2 2 "versicolor"
"83" 2 1 2 2 "versicolor"
"84" 2 1 3 2 "versicolor"
"85" 1 2 2 2 "versicolor"
"86" 2 3 2 2 "versicolor"
"87" 3 2 2 2 "versicolor"
"88" 3 1 2 2 "versicolor"
"89" 2 2 2 2 "versicolor"
"90" 1 1 2 2 "versicolor"
"91" 1 1 2 2 "versicolor"
"92" 2 2 2 2 "versicolor"
"93" 2 1 2 2 "versicolor"
"94" 1 1 2 2 "versicolor"
"95" 2 1 2 2 "versicolor"
"96" 2 2 2 2 "versicolor"
"97" 2 1 2 2 "versicolor"
"98" 3 1 2 2 "versicolor"
"99" 1 1 2 2 "versicolor"
"100" 2 1 2 2 "versicolor"
"101" 3 2 3 3 "virginica"
"102" 2 1 3 3 "virginica"
"103" 3 2 3 3 "virginica"
"104" 3 1 3 3 "virginica"
"105" 3 2 3 3 "virginica"
"106" 3 2 3 3 "virginica"
"107" 1 1 2 2 "virginica"
"108" 3 1 3 3 "virginica"
"109" 3 1 3 3 "virginica"
"110" 3 3 3 3 "virginica"
"111" 3 2 3 3 "virginica"
"112" 3 1 3 3 "virginica"
"113" 3 2 3 3 "virginica"
"114" 2 1 3 3 "virginica"
"115" 2 1 3 3 "virginica"
"116" 3 2 3 3 "virginica"
"117" 3 2 3 3 "virginica"
"118" 3 3 3 3 "virginica"
"119" 3 1 3 3 "virginica"
"120" 2 1 3 2 "virginica"
"121" 3 2 3 3 "virginica"
"122" 2 1 3 3 "virginica"
"123" 3 1 3 3 "virginica"
"124" 3 1 3 3 "virginica"
"125" 3 2 3 3 "virginica"
"126" 3 2 3 3 "virginica"
"127" 3 1 3 3 "virginica"
"128" 2 2 3 3 "virginica"
"129" 3 1 3 3 "virginica"
"130" 3 2 3 2 "virginica"
"131" 3 1 3 3 "virginica"
"132" 3 3 3 3 "virginica"
"133" 3 1 3 3 "virginica"
"134" 3 1 3 2 "virginica"
"135" 2 1 3 2 "virginica"
"136" 3 2 3 3 "virginica"
"137" 3 3 3 3 "virginica"
"138" 3 2 3 3 "virginica"
"139" 2 2 3 3 "virginica"
"140" 3 2 3 3 "virginica"
"141" 3 2 3 3 "virginica"
"142" 3 2 3 3 "virginica"
"143" 2 1 3 3 "virginica"
"144" 3 2 3 3 "virginica"
"145" 3 2 3 3 "virginica"
"146" 3 2 3 3 "virginica"
"147" 3 1 3 3 "virginica"
"148" 3 2 3 3 "virginica"
"149" 3 3 3 3 "virginica"
"150" 2 2 3 3 "virginica"