biz.sc.mm.dm.modelstats.tree

Interface Node

All Superinterfaces:

Verifiable

All Known Subinterfaces:

ClassificationNode, RegressionNode

public interface Node
extends Verifiable

Data structure for a single general tree node. Contains some node statistics and data statistics. Contains also information about the node split.

Method Summary

Methods

Modifier and Type	Method and Description
int	getActualSplitIndex() Returns index of the split.
java.util.List	getBestSplits() List of best splits.
int	getBuiltSplitIndex() Returns index of the split that was used by build task.
java.util.List	getChildren() Gets the list of this node children.
int	getChildrenCount() Gets the number of children.
double	getCount() Gets the total number of observations that fall into this node.
java.lang.String	getName() Gets the name of this node.
int	getNodeID() Gets the unique id of this node.
java.lang.String	getNodeRule() Returns string representation of the rule for this node.
double	getNullCount() Gets the number of observations that fall into this node with null value at the split attribute.
Split	getSplit() Returns actual split.
Split	getSplit(int spIndex) Returns [spIndex] best split.
java.lang.String	getSplitAttributeName() Gets the name of the split attribute.
double	getValidationCount() Gets the number of observations from validation data set that fall into this node.
boolean	isDirty() Check if this node is dirty.
boolean	isLeaf() Check if this node is a leaf node.
boolean	isPruned() Check if this node was pruned.
void	setActualSplitIndex(int i) Sets the actual split, for details see - getActualSplitIndex().
void	setBestSplits(java.util.List bestSplits) List of best splits.
void	setCount(double count) Sets the total number of observations that fall into this node.
void	setDirty(boolean dirty) Sets whether this node is dirty.
void	setLeaf(boolean isLeaf) Sets whether this node is a leaf node.
void	setNodeID(int id) Sets the unique id of this node.
void	setNullCount(double nullCount) Sets the number of observations that fall into this node with null value at the split attribute.
void	setValidationCount(double validationCount) Sets the total number of observations that fall into this node.

Methods inherited from interface biz.sc.mm.dm.base.Verifiable

verify

Method Detail

getChildren

java.util.List getChildren()

Gets the list of this node children.

Returns:

List of children. The list contains elements of Node type.

getCount

double getCount()

Gets the total number of observations that fall into this node. If there are missing values in the training data set, this can be real number.

Returns:

Number of observations that fall into this node.

setCount

void setCount(double count)

Sets the total number of observations that fall into this node. If there are missing values in the training data set, this can be real number.

Parameters:

count - Number of observations that fall into this node.

setValidationCount

void setValidationCount(double validationCount)

Sets the total number of observations that fall into this node. If there are missing values in the validation data set, this can be real number.

Parameters:

count - Number of observations in validation set that fall into this node.

getNullCount

double getNullCount()

Gets the number of observations that fall into this node with null value at the split attribute.

Returns:

Number of observations with null value at split attribute

setNullCount

void setNullCount(double nullCount)

Sets the number of observations that fall into this node with null value at the split attribute.

Parameters:

nullCount - Number of observations with null value at split attribute

getValidationCount

double getValidationCount()

Gets the number of observations from validation data set that fall into this node. If there are missing values in the validation data set, this can be real number.

Returns:

Number of observations that fall into this node.

getName

java.lang.String getName()

Gets the name of this node. It is the string representation of unique id assigned to this node, by using getNodeID() function.

Returns:

Name of this node.

getNodeID

int getNodeID()

Gets the unique id of this node.

Returns:

unique id

setNodeID

void setNodeID(int id)

Sets the unique id of this node.

Parameters:

id - the unique id

getChildrenCount

int getChildrenCount()

Gets the number of children. Determined by using getChildren() object.

Returns:

Number of children.

getSplitAttributeName

java.lang.String getSplitAttributeName()

Gets the name of the split attribute. Dynamically created using predicate object returned by #getPredicate() function. Not used by the scoring mechanism.

Returns:

Name of the split attribute.

isLeaf

boolean isLeaf()

Check if this node is a leaf node.

Returns:

If True, this is leaf node.

setLeaf

void setLeaf(boolean isLeaf)

Sets whether this node is a leaf node.

Parameters:

isLeaf - If True, this is leaf node.

isPruned

boolean isPruned()

Check if this node was pruned.

Returns:

If True, this node was pruned.

isDirty

boolean isDirty()

Check if this node is dirty. This flag can be used by algorithm to rebuild the tree.

Returns:

If True, this node is dirty.

setDirty

void setDirty(boolean dirty)

Sets whether this node is dirty.

Parameters:

dirty - If True, this node is dirty.

getNodeRule

java.lang.String getNodeRule()

Returns string representation of the rule for this node. The rule is a conjunction of all predicates from root node to this node. Not used by scoring mechanism, serves only as information about the rule.

Returns:

String representation of the rule for this node.

getActualSplitIndex

int getActualSplitIndex()

Returns index of the split. The returned split is the actual split which currently works while using the tree for viewing, applying, testing etc. After build task, it is equal to getBuiltSplitIndex() and can be manually changed.

Returns:

index of the actual split

setActualSplitIndex

void setActualSplitIndex(int i)

Sets the actual split, for details see - getActualSplitIndex().

Parameters:

i - index of the actual split

getBuiltSplitIndex

int getBuiltSplitIndex()

Returns index of the split that was used by build task. That means the split is either a split modified manually in previous model, or optimal using split evaluation criterion.

Returns:

index of the build split

getSplit

Split getSplit()

Returns actual split. The result is identical to getBestSplits()[getActualSplitIndex()].

getSplit

Split getSplit(int spIndex)

Returns [spIndex] best split. The result is identical to getBestSplits()[spIndex].

Parameters:

spIndex -

Returns:

getBestSplits

java.util.List getBestSplits()

List of best splits.

Returns:

list of Split objects.

setBestSplits

void setBestSplits(java.util.List bestSplits)

List of best splits.