Add threshold comparison above arrow in plot_tree

I realised after #13971 and also a lecture that I gave that we don’t add “True/False”, “Yes/No” above the arrow in plot_tree. Therefore, we know the criterion for the decision and the further splits but we don’t have any idea what path a sample will take (or at least it makes it more difficult to infer it).

I think that we should annotate the arrow depending if the comparison in the node is True/False.

@amueller @NicolasHug @thomasjpfan WDYT?

[d-kleine]

I found an interesting insight to above issue:

import numpy as np
from matplotlib import pyplot as plt

from sklearn.model_selection import train_test_split
from sklearn.datasets import load_iris
from sklearn.tree import DecisionTreeClassifier

iris = load_iris()
X = iris.data
y = iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)

clf = DecisionTreeClassifier(max_leaf_nodes=3, random_state=0)
clf.fit(X_train, y_train)

When using plot_tree(), it does not show the “True”/“False” labels for the threshold:

from sklearn import tree

tree.plot_tree(clf, feature_names=iris.feature_names, class_names=iris.target_names, filled=True)
plt.show()

But the “True”/“False” labels will be printed when using export_graphviz():

import graphviz

dot_data = tree.export_graphviz(clf, 
                                out_file=None, 
                                feature_names=iris.feature_names,  
                                class_names=iris.target_names,
                                filled=True)

graphviz.Source(dot_data, format="png") 

Seems like export_graphviz() is creating “True”/“False” labels for the first condition, but plot_tree() does not.

[glemaitre]

Go for it

On Mon, 27 Jan 2020 at 14:32, Rick Mackenbach [e-mail] wrote:

I could work on this

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/scikit-learn/scikit-learn/issues/16153?email_source=notifications&email_token=ABY32P37FBSRJTLKBPA25CTQ73O6HA5CNFSM4KISEEJKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEJ7P2HY#issuecomment-578747679, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABY32P3DH65VKOQPXE6CCELQ73O6HANCNFSM4KISEEJA .

– Guillaume Lemaitre Scikit-learn @ Inria Foundation https://glemaitre.github.io/