Module: tfdf.py_tree.condition

View source on GitHub

Conditions / splits for non-leaf nodes.

A condition (e.g. a>0.5) is evaluated to a binary value (e.g. True if a=5). Condition evaluations control the branching of an example in a tree.

Classes

class AbstractCondition: Generic condition.

class CategoricalIsInCondition: Condition of the form "attribute in [...set of items...]".

class CategoricalSetContainsCondition: Condition of the form "attribute intersect [...set of items...]!=empty".

class IsMissingInCondition: Condition of the form "attribute is missing".

class IsTrueCondition: Condition of the form "attribute is true".

class NumericalHigherThanCondition: Condition of the form "attribute >= threhsold".

class NumericalSparseObliqueCondition: Condition of the form "attributes * weights >= threshold".

class SimpleColumnSpec: Simplified representation of a column spec.

Functions

column_spec_bitmap_to_items(...): Converts a mask-bitmap into a list of elements.

column_spec_items_to_bitmap(...): Converts a list of elements into a mask-bitmap.

core_condition_to_condition(...): Converts a condition from the core to python format.

set_core_node(...): Sets a core node (proto format) from a python value.