Python: decision trees

This page looks at decision trees using Python.

There are several Python packages to support GraphViz. For more on GraphViz using Python, see Python: Diagrams with GraphViz .

The Python pydotplus package will be used here. One way to install the pydotplus package is as follows.

D:\Python38\Scripts\pip3.exe install pydotplus

Use the path to your installed version of Python. Here is the Python code [#1]

import os
from PIL import Image
from graphviz import Digraph
from sklearn import tree
import pydotplus
print("Decision tree example:")
sList1 = []
tList1 = []
pos1 = 0
for b1 in range(0,1+1):
	for b2 in range(0,1+1):
		for b3 in range(0,1+1):
			list2 = [b1,b2,b3]
			sList1.append(list2)
			tList1.append(pos1)
			pos1 += 1
s1 = str(sList1)
print("{0:s}".format(s1))
s2 = str(tList1)
print("{0:s}".format(s2))
tree1 = tree.DecisionTreeClassifier()
train1 = tree1.fit(sList1,tList1)
dot1 = tree.export_graphviz(train1,None)
graph1 = pydotplus.graph_from_dot_data(dot1)
basePs1 = "./"
saveFn1 = "dt51-1-01"
saveFt1 = "png"
saveFs0 = "{0:s}{1:s}".format(basePs1,saveFn1)
saveFs1 = "{0:s}{1:s}.{2:s}".format(basePs1,saveFn1,saveFt1)
graph1.write_png(saveFs1)
savedImage1 = Image.open(saveFs1)
savedWidth1, savedHeight1= savedImage1.size
(_, _, _, _, _, _, savedSize1, _, savedTime1, _) = os.stat(saveFs1)
print("Dot image:")
print("")
print("\t{0:s} : {1:d} x {2:d} , {3:,} bytes".format(saveFs1,savedWidth1,savedHeight1,savedSize1))

Here is the output of the Python code.

Decision tree example:
[[0, 0, 0], [0, 0, 1], [0, 1, 0], [0, 1, 1], [1, 0, 0], [1, 0, 1], [1, 1, 0], [1, 1, 1]]
[0, 1, 2, 3, 4, 5, 6, 7]
Dot image:
	./dt51-1-01.png : 2056 x 497 , 45,149 bytes

Note that one can use the decision tree output to create custom tree diagrams as needed.