org.apache.mahout.classifier.df.builder
Class DefaultTreeBuilder

java.lang.Object
  extended by org.apache.mahout.classifier.df.builder.DefaultTreeBuilder
All Implemented Interfaces:
TreeBuilder

public class DefaultTreeBuilder
extends Object
implements TreeBuilder

Builds a Decision Tree
Based on the algorithm described in the "Decision Trees" tutorials by Andrew W. Moore, available at:

http://www.cs.cmu.edu/~awm/tutorials

This class can be used when the criterion variable is the categorical attribute.


Constructor Summary
DefaultTreeBuilder()
           
 
Method Summary
 Node build(Random rng, Data data)
          Builds a Decision tree using the training data
protected static int[] randomAttributes(Random rng, boolean[] selected, int m)
          Randomly selects m attributes to consider for split, excludes IGNORED and LABEL attributes
 void setM(int m)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

DefaultTreeBuilder

public DefaultTreeBuilder()
Method Detail

setM

public void setM(int m)

build

public Node build(Random rng,
                  Data data)
Description copied from interface: TreeBuilder
Builds a Decision tree using the training data

Specified by:
build in interface TreeBuilder
Parameters:
rng - random-numbers generator
data - training data
Returns:
root Node

randomAttributes

protected static int[] randomAttributes(Random rng,
                                        boolean[] selected,
                                        int m)
Randomly selects m attributes to consider for split, excludes IGNORED and LABEL attributes

Parameters:
rng - random-numbers generator
selected - attributes' state (selected or not)
m - number of attributes to choose
Returns:
list of selected attributes' indices, or null if all attributes have already been selected


Copyright © 2008–2014 The Apache Software Foundation. All rights reserved.