com.itemanalysis.psychometrics.kernel

Class KernelRegression

java.lang.Object

com.itemanalysis.psychometrics.kernel.KernelRegression

public class KernelRegression
extends java.lang.Object

Nadaraya-Watson kernel regression estimator. Data values are not stored in this object. Only the evaluation getPoints (x-values) and their corresponding y-values are stored. This class is designed to incrementally compute the regression.

Constructor Summary

Constructors

Constructor and Description
KernelRegression(KernelFunction kernel, Bandwidth bandwidth, UniformDistributionApproximation uniform) Default constructor for the class.

Method Summary

Modifier and Type	Method and Description
double	getBandwidth()
double[]	getPoints()
int	getSampleSize()
void	increment(double x, double y) Increment the estimate by x and y
void	increment(double x, double y, double weight)
double[]	value() Computes teh kernel regression estimate at teh evaluation points.

Methods inherited from class java.lang.Object

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

Constructor Detail

KernelRegression

public KernelRegression(KernelFunction kernel,
                        Bandwidth bandwidth,
                        UniformDistributionApproximation uniform)

Default constructor for the class.

Parameters:

kernel - type of kernel function

bandwidth - bandwidth

uniform - distribution representing the evaluation points

Method Detail

increment

public void increment(double x,
                      double y)

Increment the estimate by x and y

Parameters:

x - value of teh independent variable

y - value of the dependent variable

increment

public void increment(double x,
                      double y,
                      double weight)

value

public double[] value()

Computes teh kernel regression estimate at teh evaluation points.

Returns:

predicted values of y

getPoints

public double[] getPoints()

Returns:

evaluae pints points

getBandwidth

public double getBandwidth()

Returns:

bandwidth of the kernel regression estimator

getSampleSize

public int getSampleSize()

Returns:

sample size use din teh computation of teh regression