For the paper, it would be really useful to visualise the output tree. Apparently this ias actually really easy with scikit-learn:
from sklearn.tree import DecisionTreeClassifier
from sklearn import tree
from matplotlib import pyplot as plt
dataset=np.concatenate([mat_original,np.expand_dims(y_km,axis=1)],axis=1)
dt = DecisionTreeClassifier(max_depth=15, min_samples_leaf=30)
classifier=dt.fit(dataset[:,:-1], dataset[:,-1])
fig, axes=plt.subplots(1, 1, figsize=(400, 200), sharex=True, sharey=True)
tree.plot_tree(classifier)
fig.savefig('tree.png')
It would be interesting to combine this code with the idea to make visualisation in a notebook easy. We could create a wrapper of the class that has a _to_html() method so that when you execute a cell with just the tree object name, it prints out a graph, like the nice options you get with iris cubes or pandas dataframes in a jupyter notebook. This could potentially be fed back to scikit learn in some way.
For the paper, it would be really useful to visualise the output tree. Apparently this ias actually really easy with scikit-learn:
It would be interesting to combine this code with the idea to make visualisation in a notebook easy. We could create a wrapper of the class that has a
_to_html()
method so that when you execute a cell with just the tree object name, it prints out a graph, like the nice options you get with iris cubes or pandas dataframes in a jupyter notebook. This could potentially be fed back to scikit learn in some way.