scikit-learn tree.DecisionTreeClassifier export to pdf too wide to fit

Question

I followed the the instructions in this thread to get my export working Python, PyDot and DecisionTree but a pdf export truncates the tree after a page width and the png is too small to read. Is there a way of creating a pdf with unlimited width?

My code is

from sklearn import tree

clf = tree.DecisionTreeClassifier().fit(X_train, y_train)


import pydotplus
from IPython.display import Image
from sklearn.externals.six import StringIO
dot_data = StringIO()
tree.export_graphviz(clf, out_file=dot_data,feature_names=cols_scaled.columns) 
graph = pydotplus.graph_from_dot_data(dot_data.getvalue()) 
Image(graph.create_png())
graph.write_pdf("decisiontree.pdf")

Many thanks

For the png, there must be some parameter which controls the dimensions — Vivek Kumar, Jun 08 '17 at 01:47
How do I make it so that the PNG isn't made inline in the console window? I'm using Windows 10, Python 3.5 and Spyder — jlt199, Jun 08 '17 at 15:17

score 0 · Answer 1 · edited May 13 '18 at 20:35

Try using the following code, PDF requires width and let me know if it worked.

from sklearn.externals.six import StringIO
import pydotplus   #using pydotplus in windows10, python 3.6.X
dot_data = StringIO()

tree.export_graphviz(clf, out_file=dot_data, 
                         feature_names=cols_scaled.column,  
                         class_names=<replace target column here>,  
                         filled=True, rounded=True,  
                         special_characters=True)  
graph = pydotplus.graph_from_dot_data(dot_data.getvalue())  
graph.write_pdf("iris.pdf")

scikit-learn tree.DecisionTreeClassifier export to pdf too wide to fit

1 Answers1