Package ciir.umass.edu.learning.tree

Class RegressionTree

java.lang.Object

ciir.umass.edu.learning.tree.RegressionTree

public class RegressionTree
extends java.lang.Object

Author:

vdang

Field Summary

Fields

Modifier and Type	Field	Description
protected int[]	features
protected FeatureHistogram	hist
protected int[]	index
protected java.util.List<Split>	leaves
protected int	minLeafSupport
protected int	nodes
protected Split	root
protected float[][]	thresholds
protected double[]	trainingLabels
protected DataPoint[]	trainingSamples

Constructor Summary

Constructors

Constructor	Description
RegressionTree(int nLeaves, DataPoint[] trainingSamples, double[] labels, FeatureHistogram hist, int minLeafSupport)
RegressionTree(Split root)

Method Summary

Modifier and Type	Method	Description
void	clearSamples()	Clear samples associated with each leaves (when they are no longer necessary) in order to save memory
double	eval(DataPoint dp)	Get the tree output for the input sample
void	fit()	Fit the tree from the specified training data
protected void	insert(java.util.List<Split> ls, Split s)
java.util.List<Split>	leaves()	Retrieve all leave nodes in the tree
java.lang.String	toString()	Generate the string representation of the tree
java.lang.String	toString(java.lang.String indent)
double	variance()

Methods inherited from class java.lang.Object

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

Field Detail

nodes

protected int nodes

minLeafSupport

protected int minLeafSupport

root

protected Split root

leaves

protected java.util.List<Split> leaves

trainingSamples

protected DataPoint[] trainingSamples

trainingLabels

protected double[] trainingLabels

features

protected int[] features

thresholds

protected float[][] thresholds

index

protected int[] index

hist

protected FeatureHistogram hist

Constructor Detail

RegressionTree

public RegressionTree(Split root)

RegressionTree

public RegressionTree(int nLeaves,
                      DataPoint[] trainingSamples,
                      double[] labels,
                      FeatureHistogram hist,
                      int minLeafSupport)

Method Detail

fit

public void fit()

Fit the tree from the specified training data

eval

public double eval(DataPoint dp)

Get the tree output for the input sample

Parameters:

dp -

Returns:

leaves

public java.util.List<Split> leaves()

Retrieve all leave nodes in the tree

Returns:

clearSamples

public void clearSamples()

Clear samples associated with each leaves (when they are no longer necessary) in order to save memory

toString

public java.lang.String toString()

Generate the string representation of the tree

Overrides:

toString in class java.lang.Object

toString

public java.lang.String toString(java.lang.String indent)

variance

public double variance()

insert

protected void insert(java.util.List<Split> ls,
                      Split s)