weka.classifiers.trees.adtree
Class Splitter

java.lang.Object
  weka.classifiers.trees.adtree.Splitter

All Implemented Interfaces:

java.lang.Cloneable, java.io.Serializable

Direct Known Subclasses:

TwoWayNominalSplit, TwoWayNumericSplit

public abstract class Splitter

extends java.lang.Object

implements java.io.Serializable, java.lang.Cloneable

Abstract class representing a splitter node in an alternating tree.

Version:

$Revision: 1.2 $

Author:

Richard Kirkby (rkirkby@cs.waikato.ac.nz)

See Also:

Serialized Form

Field Summary

int	orderAdded The number this node was in the order of nodes added to the tree

Constructor Summary

Splitter()

Method Summary

abstract java.lang.String	attributeString(Instances dataset) Gets the string describing the attributes the split depends on.
abstract int	branchInstanceGoesDown(Instance i) Gets the index of the branch that an instance applies to.
abstract java.lang.Object	clone() Clones this node.
abstract java.lang.String	comparisonString(int branchNum, Instances dataset) Gets the string describing the comparision the split depends on for a particular branch. i.e. the right hand side of the description of the split.
abstract boolean	equalTo(Splitter compare) Tests whether two splitters are equivalent.
abstract PredictionNode	getChildForBranch(int branchNum) Gets the child for a branch of the split.
abstract int	getNumOfBranches() Gets the number of branches of the split.
abstract ReferenceInstances	instancesDownBranch(int branch, Instances sourceInstances) Gets the subset of instances that apply to a particluar branch of the split.
abstract void	setChildForBranch(int branchNum, PredictionNode childPredictor) Sets the child for a branch of the split.

Methods inherited from class java.lang.Object

equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

orderAdded

public int orderAdded

The number this node was in the order of nodes added to the tree

Constructor Detail

Splitter

public Splitter()

Method Detail

getNumOfBranches

public abstract int getNumOfBranches()

Gets the number of branches of the split.

Returns:

the number of branches

branchInstanceGoesDown

public abstract int branchInstanceGoesDown(Instance i)

Gets the index of the branch that an instance applies to. Returns -1 if no branches apply.

Parameters:

i - the instance

Returns:

the branch index

instancesDownBranch

public abstract ReferenceInstances instancesDownBranch(int branch,
                                                       Instances sourceInstances)

Gets the subset of instances that apply to a particluar branch of the split. If the branch index is -1, the subset will consist of those instances that don't apply to any branch.

Parameters:

branch - the index of the branch

sourceInstances - the instances from which to find the subset

Returns:

the set of instances that apply

attributeString

public abstract java.lang.String attributeString(Instances dataset)

Gets the string describing the attributes the split depends on. i.e. the left hand side of the description of the split.

Parameters:

dataset - the dataset that the split is based on

Returns:

a string describing the attributes

comparisonString

public abstract java.lang.String comparisonString(int branchNum,
                                                  Instances dataset)

Gets the string describing the comparision the split depends on for a particular branch. i.e. the right hand side of the description of the split.

Parameters:

branchNum - the branch of the split

dataset - the dataset that the split is based on

Returns:

a string describing the comparison

equalTo

public abstract boolean equalTo(Splitter compare)

Tests whether two splitters are equivalent.

Parameters:

compare - the splitter to compare with

Returns:

whether or not they match

setChildForBranch

public abstract void setChildForBranch(int branchNum,
                                       PredictionNode childPredictor)

Sets the child for a branch of the split.

Parameters:

branchNum - the branch to set the child for

childPredictor - the new child

getChildForBranch

public abstract PredictionNode getChildForBranch(int branchNum)

Gets the child for a branch of the split.

Parameters:

branchNum - the branch to get the child for

Returns:

the child

clone

public abstract java.lang.Object clone()

Clones this node. Performs a deep copy, recursing through the tree.

Returns:

a clone