Package gorsat.Regression

Class LogisticRegression

java.lang.Object

gorsat.Regression.LogisticRegression

public abstract class LogisticRegression
extends java.lang.Object

This program fits a logistic regression model on a given data set. We find the maximum likelihood estimators by using Newton's method to find the zero of the gradient of the likelihood function.

Field Summary

Fields

Modifier and Type	Field	Description
double[]	beta
protected double[]	delta
protected int	dim
protected double[]	gradient
protected double[]	gradMult
protected double[][]	hessian
protected int	maximumNumberOfSamples
protected int	numberOfDepVars
int	numberOfIterations
protected int	numberOfSamples
protected double[]	oneOver
protected double[]	prod
double[]	pValues
protected SigmoidFunction	sf
double[]	zStats

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	LogisticRegression(int numberOfDepVars, int maximumNumberOfSamples)

Method Summary

Modifier and Type	Method	Description
protected abstract void	computeTestStats()
static LogisticRegression	getGorLogisticRegressionObject(int numberOfDepVars, int maximumNumberOfSamples)
boolean	runRegression(double[][] x, boolean[] y, int numberOfSamples, double tol, int maxIter)
protected abstract void	updateGradientAndHessian(double[][] x, boolean[] y)
protected abstract void	updateGradientAndHessianFirstRun(double[][] x, boolean[] y)

Methods inherited from class java.lang.Object

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

Field Detail

numberOfDepVars

protected final int numberOfDepVars

maximumNumberOfSamples

protected final int maximumNumberOfSamples

dim

protected final int dim

beta

public double[] beta

zStats

public double[] zStats

pValues

public double[] pValues

numberOfIterations

public int numberOfIterations

gradient

protected final double[] gradient

hessian

protected final double[][] hessian

delta

protected final double[] delta

numberOfSamples

protected int numberOfSamples

oneOver

protected final double[] oneOver

gradMult

protected final double[] gradMult

prod

protected final double[] prod

sf

protected final SigmoidFunction sf

Constructor Detail

LogisticRegression

protected LogisticRegression(int numberOfDepVars,
                             int maximumNumberOfSamples)

Method Detail

getGorLogisticRegressionObject

public static LogisticRegression getGorLogisticRegressionObject(int numberOfDepVars,
                                                                int maximumNumberOfSamples)

runRegression

public boolean runRegression(double[][] x,
                             boolean[] y,
                             int numberOfSamples,
                             double tol,
                             int maxIter)

computeTestStats

protected abstract void computeTestStats()

updateGradientAndHessian

protected abstract void updateGradientAndHessian(double[][] x,
                                                 boolean[] y)

updateGradientAndHessianFirstRun

protected abstract void updateGradientAndHessianFirstRun(double[][] x,
                                                         boolean[] y)