Visualising the decision tree in sklearn

Question

When I want to visualise the tree I got this error.

I have shown the required libraries imported. Is there expected reason with jupiter-notebook ?

from sklearn import tree
import matplotlib.pyplot
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_breast_cancer
cancer=load_breast_cancer()
x=cancer.data
y=cancer.target
clf=DecisionTreeClassifier(max_depth=1000)
x_train,x_test,y_train,y_test=train_test_split(x,y)
clf=clf.fit(x_train,y_train)
tree.plot_tree(clf.fit(x_train,y_train))

AttributeError: module 'sklearn.tree' has no attribute 'plot_tree'

Solution 1

I assigned the tree to an object and added plt.show(). This works for me.

%matplotlib inline
from sklearn import tree
import matplotlib.pyplot as plt
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
from sklearn.datasets import load_breast_cancer
cancer = load_breast_cancer()
x = cancer.data
y = cancer.target
clf = DecisionTreeClassifier(max_depth = 1000)
x_train,x_test,y_train,y_test = train_test_split(x,y)

fig = clf.fit(x_train,y_train)
tree.plot_tree(fig)
plt.show()

But I recommend using graphviz, it's much more flexible.

Solution 2

upgrade sklearn package:

pip install --upgrade sklearn

Solution 3

This is because plot_tree is new is sklearn version 0.21, as indicated in the documentation. Check if you have a version sufficient by running this:

import sklearn

print(sklearn.__version__)

assert float(sklearn.__version__[2:]) >= 21, 'sklearn version insufficient.'

If you get an error message, you need to update sklearn

pip install --upgrade sklearn

Comments

[PV8]

check it out here: scikit-learn.org/stable/modules/tree.html

[Anna Yashina]

make sure your matplotlib version is >= 1.5 and try to save your fit to an object before putting it in the plot function

[Roshan]

matplotlib version is 3.0.3

[Josmoor98]

plot_tree is new in version 0.21. Maybe check your scikit-learn version

[Roshan]

Thanks a lot, it was the version of my sklearn

[Anna Yashina]

current version is 0.21.3 you can upgrade running !pip install -U scikit-learn or conda upgrade scikit-learn if you are using anaconda

[Roshan]

Thank you so much, I have got it after upgrading

[Nicolas Gervais]

This is not an answer. You are just saying "I don't have this problem".