Gets the best split suggestion given a criterion and a class distribution
Gets the best split suggestion given a criterion and a class distribution
the split criterion to use
the class distribution before the split
the value of the feature
true to use binary splits
suggestion of best feature split
Merge the FeatureClassObserver to current FeatureClassObserver
Merge the FeatureClassObserver to current FeatureClassObserver
the FeatureClassObserver will be merged
whether called when a Hoeffding tree try to split
current FeatureClassObserver
Updates statistics of this observer given a feature value, a class index and the weight of the example observed
Updates statistics of this observer given a feature value, a class index and the weight of the example observed
the index of class
the value of the feature
the weight of the example
Not yet supported.
Not yet supported.
Gets the probability for an attribute value given a class
Gets the probability for an attribute value given a class
the index of class
the value of the feature
probability for a feature value given a class
NullFeatureClassObserver is a null class for observers.