edu.stanford.nlp.classify
Class LogPrior

java.lang.Object
  extended by edu.stanford.nlp.classify.LogPrior
All Implemented Interfaces:
java.io.Serializable

public class LogPrior
extends java.lang.Object
implements java.io.Serializable

A Prior for functions. Immutable.

Author:
Galen Andrew
See Also:
Serialized Form

Nested Class Summary
static class LogPrior.LogPriorType
           
 
Field Summary
 LogPrior.LogPriorType type
           
 
Constructor Summary
LogPrior()
           
LogPrior(double[] C)
          IMPORTANT NOTE: This constructor allows non-uniform regularization, but it transforms the inputs C (like the machine learning people like) to sigma (like we NLP folks like).
LogPrior(int intPrior)
           
LogPrior(int intPrior, double sigma, double epsilon)
           
LogPrior(LogPrior.LogPriorType type)
           
LogPrior(LogPrior.LogPriorType type, double sigma, double epsilon)
           
 
Method Summary
 double compute(double[] x, double[] grad)
          Adjust the given grad array by adding the prior's gradient component and return the value of the logPrior
 double computeStochastic(double[] x, double[] grad, double fractionOfData)
           
static LogPrior getAdaptationPrior(double[] means, LogPrior otherPrior)
           
 double getEpsilon()
           
 double getSigma()
           
 double getSigmaSquared()
           
 double[] getSigmaSquaredM()
           
 LogPrior.LogPriorType getType()
           
static LogPrior.LogPriorType getType(java.lang.String name)
           
 void setEpsilon(double epsilon)
           
 void setSigma(double sigma)
           
 void setSigmaSquared(double sigmaSq)
           
 void setSigmaSquaredM(double[] sigmaSq)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

type

public final LogPrior.LogPriorType type
Constructor Detail

LogPrior

public LogPrior()

LogPrior

public LogPrior(int intPrior)

LogPrior

public LogPrior(LogPrior.LogPriorType type)

LogPrior

public LogPrior(int intPrior,
                double sigma,
                double epsilon)

LogPrior

public LogPrior(LogPrior.LogPriorType type,
                double sigma,
                double epsilon)

LogPrior

public LogPrior(double[] C)
IMPORTANT NOTE: This constructor allows non-uniform regularization, but it transforms the inputs C (like the machine learning people like) to sigma (like we NLP folks like). C = 1/\sigma^2

Method Detail

getType

public static LogPrior.LogPriorType getType(java.lang.String name)

getAdaptationPrior

public static LogPrior getAdaptationPrior(double[] means,
                                          LogPrior otherPrior)

getType

public LogPrior.LogPriorType getType()

getSigma

public double getSigma()

getSigmaSquared

public double getSigmaSquared()

getSigmaSquaredM

public double[] getSigmaSquaredM()

getEpsilon

public double getEpsilon()

setSigma

public void setSigma(double sigma)

setSigmaSquared

public void setSigmaSquared(double sigmaSq)

setSigmaSquaredM

public void setSigmaSquaredM(double[] sigmaSq)

setEpsilon

public void setEpsilon(double epsilon)

computeStochastic

public double computeStochastic(double[] x,
                                double[] grad,
                                double fractionOfData)

compute

public double compute(double[] x,
                      double[] grad)
Adjust the given grad array by adding the prior's gradient component and return the value of the logPrior

Parameters:
x - the input point
grad - the gradient array
Returns:
the value


Stanford NLP Group