com.numericalmethod.suanshu.stats.regression.linear.glm.distribution.link

Class Logit

java.lang.Object

com.numericalmethod.suanshu.stats.regression.linear.glm.distribution.link.Logit

All Implemented Interfaces:

LinkFunction, java.io.Serializable

public class Logit
extends java.lang.Object
implements LinkFunction

This class represents the link function:

               mu
 g(x) = log(--------)
             1 - mu
 

See Also:

GeneralizedLinearModel, Serialized Form

Constructor Summary

Constructors

Constructor and Description
Logit()

Method Summary

Modifier and Type	Method and Description
double	derivative(double x) Derivative of the link function, i.e., g'(x).
double	inverse(double x) Inverse of the link function, i.e., g-1(x).

Methods inherited from class java.lang.Object

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

Constructor Detail

Logit

public Logit()

Method Detail

inverse

public double inverse(double x)

Inverse of the link function, i.e., g-1(x).

               exp(x)
 g_inv(x) = ------------
             1 + exp(x)
 

Specified by:

inverse in interface LinkFunction

Parameters:

x - x

Returns:

g^-1(x)

derivative

public double derivative(double x)

Derivative of the link function, i.e., g'(x).

              1
 g'(x) = -----------
          x (1 - x)
 

Specified by:

derivative in interface LinkFunction

Parameters:

x - x

Returns:

g'(x)