chriswbartley / monoensemble Goto Github PK

sklearn.utils.validation.check_is_fitted no longer takes the second argument (e.g. 'prior' in mono_gradient_boosting.py lines 130, 156.

So, after a few days double-testing the updated version, after fitting any model, for example the RF one from the official docs, I get the following error raised if I try to access the fitted model's feature_importances_:

Traceback (most recent call last):
File "/home/vasselai/.local/lib/python3.6/site-packages/joblib/parallel.py", line 820, in dispatch_one_
batch
tasks = self._ready_batches.get(block=False)
File "/usr/lib/python3.6/queue.py", line 161, in get
raise Empty
queue.Empty

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "", line 1, in
File "/home/vasselai/.local/lib/python3.6/site-packages/sklearn/ensemble/forest.py", line 450, in feat
ure_importances
for tree in self.estimators_ if tree.tree_.node_count > 1)
File "/home/vasselai/.local/lib/python3.6/site-packages/joblib/parallel.py", line 1041, in call
if self.dispatch_one_batch(iterator):
File "/home/vasselai/.local/lib/python3.6/site-packages/joblib/parallel.py", line 831, in dispatch_one_
batch
islice = list(itertools.islice(iterator, big_batch_size))
File "/home/vasselai/.local/lib/python3.6/site-packages/sklearn/ensemble/forest.py", line 450, in
for tree in self.estimators if tree.tree_.node_count > 1)
AttributeError: 'MonoGradientBoostingClassifier' object has no attribute 'tree_'

At first I was under the impression that the monoensemble code was just not storing the internal fitted tree object inside an internal tree_, but that did not solve it. There's something else at play. Given the criticality of being able to explore feature importances, I thought best to bring this up officially.

for multi-class case with n-classes the predict_proba method seems to return (n-1) probabilities. The probabilities also seem to be un-normalized so it is not possible to recover the probability of the missing class using the probabilities of the rest of the classes

Currently predict_prob gives [P(y>1), P(y>2), P(y>3)] cumulative probabilities, consider calculating class prpobabilities instead, and making these probabilities available as predict_cum_proba().