I'm using a histogram values as features in the machine learning task.
How to select the best bins?
I thought at the beginning to break the range into large bins, than break significant bins again but until when?
And may be there a special methods for histogram bins selection in machine learning case?