/*
* File: RegressionTreeNode.java
* Authors: Justin Basilico
* Company: Sandia National Laboratories
* Project: Cognitive Foundry
*
* Copyright November 29, 2007, Sandia Corporation. Under the terms of Contract
* DE-AC04-94AL85000, there is a non-exclusive license for use of this work by
* or on behalf of the U.S. Government. Export of this program may require a
* license from the United States Government. See CopyrightHistory.txt for
* complete details.
*
*/
package gov.sandia.cognition.learning.algorithm.tree;
import gov.sandia.cognition.evaluator.Evaluator;
import gov.sandia.cognition.learning.function.categorization.Categorizer;
import gov.sandia.cognition.util.ObjectUtil;
/**
* The {@code RegressionTreeNode} implements a {@code DecisionTreeNode} for
* a tree that does regression.
*
* @param <InputType> The input type for the tree.
* @param <InteriorType> The type that is the output of the decision made at
* this tree node.
* @author Justin Basilico
* @since 2.0
*/
public class RegressionTreeNode<InputType, InteriorType>
extends AbstractDecisionTreeNode<InputType, Double, InteriorType>
{
/** The default value for the node is 0.0. */
public static final double DEFAULT_VALUE = 0.0;
/** The function to apply for leaf nodes. */
protected Evaluator<? super InputType, Double> scalarFunction;
/** The value stored at the tree node. It is used as a backup value if
* no scalar function exists for the node but the output is requested. */
protected double value;
/**
* Creates a new instance of RegressionTreeNode.
*/
public RegressionTreeNode()
{
this(null, DEFAULT_VALUE);
}
/**
* Creates a new instance of RegressionTreeNode.
*
* @param parent The parent node of this node. Null if this is a root.
* @param value The value stored at the node.
*/
public RegressionTreeNode(
final DecisionTreeNode<InputType, Double> parent,
final double value)
{
this(parent, null, null, value, null);
}
/**
* Creates a new instance of RegressionTreeNode.
*
* @param parent The parent node of this node. Null if this is a root.
* @param decider The decision function for interior nodes.
* @param value The value stored at the node.
*/
public RegressionTreeNode(
final DecisionTreeNode<InputType, Double> parent,
final Categorizer<? super InputType, ? extends InteriorType> decider,
final double value)
{
this(parent, decider, null, value, null);
}
/**
* Creates a new instance of RegressionTreeNode.
*
* @param parent The parent node of this node. Null if this is a root.
* @param scalarFunction The scalar function to apply at leaf nodes.
* @param value The value stored at the node.
*/
public RegressionTreeNode(
final DecisionTreeNode<InputType, Double> parent,
final Evaluator<? super InputType, Double> scalarFunction,
final double value)
{
this(parent, scalarFunction, value, null);
}
/**
* Creates a new instance of RegressionTreeNode.
*
* @param parent The parent node of this node. Null if this is a root.
* @param scalarFunction The scalar function to apply at leaf nodes.
* @param value The value stored at the node.
* @param incomingValue The incoming value.
*/
public RegressionTreeNode(
final DecisionTreeNode<InputType, Double> parent,
final Evaluator<? super InputType, Double> scalarFunction,
final double value,
final Object incomingValue)
{
this(parent, null, scalarFunction, value, incomingValue);
}
/**
* Creates a new instance of RegressionTreeNode.
*
* @param parent The parent node of this node. Null if this is a root.
* @param decider The decision function for interior nodes.
* @param scalarFunction The scalar function to apply at leaf nodes.
* @param value The value stored at the node.
* @param incomingValue The incoming value.
*/
public RegressionTreeNode(
final DecisionTreeNode<InputType, Double> parent,
final Categorizer<? super InputType, ? extends InteriorType> decider,
final Evaluator<? super InputType, Double> scalarFunction,
final double value,
final Object incomingValue)
{
super(parent, decider, incomingValue);
this.setScalarFunction(scalarFunction);
this.setValue(value);
}
@Override
@SuppressWarnings("unchecked")
public RegressionTreeNode<InputType, InteriorType>
clone()
{
RegressionTreeNode<InputType, InteriorType> result = (RegressionTreeNode<InputType, InteriorType>)
super.clone();
result.scalarFunction = ObjectUtil.cloneSmart(this.scalarFunction);
return result;
}
public Double getOutput(
final InputType input)
{
if ( this.scalarFunction == null )
{
// No function, so output the value at the node.
return value;
}
else
{
// Evaluate the function on the output.
return this.scalarFunction.evaluate(input);
}
}
/**
* Gets the scalar function applied to the input when the node is a leaf
* node.
*
* @return The scalar function applied to the input for leaves.
*/
public Evaluator<? super InputType, Double> getScalarFunction()
{
return this.scalarFunction;
}
/**
* Sets the scalar function applied to the input when the node is a leaf
* node.
*
* @param scalarFunction
* The scalar function applied to the input for leaves.
*/
public void setScalarFunction(
final Evaluator<? super InputType, Double> scalarFunction)
{
this.scalarFunction = scalarFunction;
}
/**
* Gets the value stored at the node, which is usually the mean value.
*
* @return The value stored at the node.
*/
public double getValue()
{
return this.value;
}
/**
* Sets the value stored at the node, which is usually the mean value.
*
* @param value
* The value stored at the node.
*/
public void setValue(
final double value)
{
this.value = value;
}
}