Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Multinomial Bayes issue #5814
Given the digits dataset (available in sklearn.datasets), we split it into train and test set.
It seems the BernoulliNB class from sklearn also has this problem.
My inefficient but straightforward implementation performs expectedly well on the training set.
Since it's my first time submitting an error report I'm completely lost :(
It says here that PR-s are used to tell others about changes I pushed to a repository. I didn't push anything O_o, nor do I want to. The demo script I submitted (via pastebin link) doesn't fix the issue, or change anything in sklearn code. It's just there to reproduce the issue, so someone who knows the sklearn code can debug it more easily. Perhaps I'm misunderstanding something.
It seems when