Tree Implementations

Tree

 Tree ()

Tree class enable building all simpe single tree models

source

CARTRegressionTree

 CARTRegressionTree (objective:str='class', min_samples_split:int=20,
                     min_sample_leaf:int=10, max_compete:int=4,
                     min_impurity:float=0.01, max_depth:int=inf,
                     max_surrogates:int=2, surrogate_style:int=2,
                     parallelize:str='feature', tree_growth:str='cart',
                     feature_weights:Union[float,int,list]=None,
                     loss:Union[str,float]=None,
                     verbose:Union[bool,int]=False,
                     digits:Optional[int]=2)

Super class of all kinds of tree.

	Type	Default	Details
objective	str	class	defining objective of the whole tree building.
min_samples_split	int	20	The minimum number of samples needed to make a split when building a tree.
min_sample_leaf	int	10	Minimum sample required to have a leaf node
max_compete	int	4	the number of competitor splits retained in the output. It is useful to know not just which split was chosen, but which variable came in second, third, etc.
min_impurity	float	0.01	The minimum impurity required to split the tree further. this is equivalent to complexity in CART
max_depth	int	inf	The maximum depth of a tree.
max_surrogates	int	2	more on this to make it generalized and not CART Specific
surrogate_style	int	2	more on this to make it generalized and not CART Specific
parallelize	str	feature	follow form LightGBMs behavior
tree_growth	str	cart	follow binary structure if cart else follow multiple structure as C4.5
feature_weights	Union	None	weight of each feature in the split. Default is set to 1 for all features
loss	Union	None	String of loss or function which defines the loss. This amounts to Loss function that is used for Gradient Boosting models to calculate impurity.
verbose	Union	False	Verbosity for tree building
digits	Optional	2	To round the values before doing a split

source

CART

 CART (objective:str='class', min_samples_split:int=20,
       min_sample_leaf:int=10, max_compete:int=4, min_impurity:float=0.01,
       max_depth:int=inf, max_surrogates:int=2, surrogate_style:int=2,
       parallelize:str='feature', tree_growth:str='cart',
       feature_weights:Union[float,int,list]=None,
       loss:Union[str,float]=None, verbose:Union[bool,int]=False,
       digits:Optional[int]=2)

Super class of all kinds of tree.

	Type	Default	Details
objective	str	class	defining objective of the whole tree building.
min_samples_split	int	20	The minimum number of samples needed to make a split when building a tree.
min_sample_leaf	int	10	Minimum sample required to have a leaf node
max_compete	int	4	the number of competitor splits retained in the output. It is useful to know not just which split was chosen, but which variable came in second, third, etc.
min_impurity	float	0.01	The minimum impurity required to split the tree further. this is equivalent to complexity in CART
max_depth	int	inf	The maximum depth of a tree.
max_surrogates	int	2	more on this to make it generalized and not CART Specific
surrogate_style	int	2	more on this to make it generalized and not CART Specific
parallelize	str	feature	follow form LightGBMs behavior
tree_growth	str	cart	follow binary structure if cart else follow multiple structure as C4.5
feature_weights	Union	None	weight of each feature in the split. Default is set to 1 for all features
loss	Union	None	String of loss or function which defines the loss. This amounts to Loss function that is used for Gradient Boosting models to calculate impurity.
verbose	Union	False	Verbosity for tree building
digits	Optional	2	To round the values before doing a split

source

C45

 C45 (objective:str='class', min_samples_split:int=20,
      min_sample_leaf:int=10, max_compete:int=4, min_impurity:float=0.01,
      max_depth:int=inf, max_surrogates:int=2, surrogate_style:int=2,
      parallelize:str='feature', tree_growth:str='cart',
      feature_weights:Union[float,int,list]=None,
      loss:Union[str,float]=None, verbose:Union[bool,int]=False,
      digits:Optional[int]=2)

Super class of all kinds of tree.

	Type	Default	Details
objective	str	class	defining objective of the whole tree building.
min_samples_split	int	20	The minimum number of samples needed to make a split when building a tree.
min_sample_leaf	int	10	Minimum sample required to have a leaf node
max_compete	int	4	the number of competitor splits retained in the output. It is useful to know not just which split was chosen, but which variable came in second, third, etc.
min_impurity	float	0.01	The minimum impurity required to split the tree further. this is equivalent to complexity in CART
max_depth	int	inf	The maximum depth of a tree.
max_surrogates	int	2	more on this to make it generalized and not CART Specific
surrogate_style	int	2	more on this to make it generalized and not CART Specific
parallelize	str	feature	follow form LightGBMs behavior
tree_growth	str	cart	follow binary structure if cart else follow multiple structure as C4.5
feature_weights	Union	None	weight of each feature in the split. Default is set to 1 for all features
loss	Union	None	String of loss or function which defines the loss. This amounts to Loss function that is used for Gradient Boosting models to calculate impurity.
verbose	Union	False	Verbosity for tree building
digits	Optional	2	To round the values before doing a split