Split

class bartpy.split.SplitCondition(splitting_variable: str, splitting_value: float)

A representation of a split in feature space. The two main components are:

• splitting_variable: which variable is being split on

• splitting_value: the value being split on

  all values less than or equal to this go left, all values greater go right

condition(data: bartpy.data.Data, cached=True) → numpy.ndarray

Returns a Bool array indicating which side of the split each row of Data should go False => Left True => Right

left(data: bartpy.data.Data) → typing.Tuple[_ForwardRef('SplitCondition'), numpy.ndarray]

Returns a Bool array indicating whether each row should go into the left split. Inverse of self.right

right(data: bartpy.data.Data) → typing.Tuple[_ForwardRef('SplitCondition'), numpy.ndarray]

Returns a Bool array indicating whether each row should go into the left split. Inverse of self.right