LBJ2.learn
Class BinaryMIRA

java.lang.Object
  LBJ2.classify.Classifier
      LBJ2.learn.Learner
          LBJ2.learn.LinearThresholdUnit
              LBJ2.learn.SparsePerceptron
                  LBJ2.learn.BinaryMIRA

All Implemented Interfaces:

java.io.Serializable, java.lang.Cloneable

public class BinaryMIRA
extends SparsePerceptron

The Binary MIRA learning algorithm implementation. This algorithm operates very similarly to SparsePerceptron with a thick separator, except the learning rate is a function of each training example's margin. When the weight vector has made a mistake, the full SparsePerceptron.learningRate will be used. When the weight vector did not make a mistake, SparsePerceptron.learningRate is multiplied by the following value before the update takes place.

(beta/2 - y(w*x)) / ||x||²

In the expression above, w is the weight vector, y represents the label of the example vector x, * stands for inner product, and beta is a user supplied parameter. If this expression turns out to be non-positive (i.e., if y(w*x) >= beta/2), no update is made for that example.

It is assumed that Learner.labeler is a single discrete classifier that produces the same feature for every example object and that the values that feature may take are available through the Classifier.allowableValues() method. The second value returned from Classifier.allowableValues() is treated as "positive", and it is assumed there are exactly 2 allowable values. Assertions will produce error messages if these assumptions do not hold.

This algorithm's user-configurable parameters are stored in member fields of this class. They may be set via either a constructor that names each parameter explicitly or a constructor that takes an instance of Parameters as input. The documentation in each member field in this class indicates the default value of the associated parameter when using the former type of constructor. The documentation of the associated member field in the Parameters class indicates the default value of the parameter when using the latter type of constructor.

See Also:

Serialized Form

Nested Class Summary

static class

BinaryMIRA.Parameters Simply a container for all of BinaryMIRA's configurable parameters.

Field Summary

protected double	beta The user supplied learning algorithm parameter; default defaultBeta.
static double	defaultBeta Default value for beta.
protected double	saveLearningRate A copy of the original learning rate value is stored here so that SparsePerceptron.learningRate can be modified without losing it.
static double	TOLERANCE Used to decide if two values are nearly equal to each other.

Fields inherited from class LBJ2.learn.SparsePerceptron

defaultLearningRate, learningRate

Fields inherited from class LBJ2.learn.LinearThresholdUnit

allowableValues, bias, defaultInitialWeight, defaultThickness, defaultThreshold, defaultWeightVector, initialWeight, negativeThickness, positiveThickness, threshold, weightVector

Fields inherited from class LBJ2.learn.Learner

extractor, labeler

Fields inherited from class LBJ2.classify.Classifier

containingPackage, name

Constructor Summary

BinaryMIRA()
The learning rate and beta take default values while the name of the classifier takes the empty string.

BinaryMIRA(BinaryMIRA.Parameters p)
Initializing constructor.

BinaryMIRA(double r)
Sets the learning rate to the specified value, and beta to the default, while the name of the classifier takes the empty string.

BinaryMIRA(double r, double B)
Sets the learning rate and beta to the specified values, while the name of the classifier takes the empty string.

BinaryMIRA(double r, double B, SparseWeightVector v)
Sets the learning rate, beta and the weight vector to the specified values.

BinaryMIRA(java.lang.String n)
Sets the name of the classifier to the specified value, while the learning rate and beta take default values.

BinaryMIRA(java.lang.String n, BinaryMIRA.Parameters p)
Initializing constructor.

BinaryMIRA(java.lang.String n, double r)
Sets the name of the classifier and learning rate to the specified values, while beta takes the default value.

BinaryMIRA(java.lang.String n, double r, double B)
Sets the name of the classifier, the learning rate and beta to the specified values.

BinaryMIRA(java.lang.String n, double r, double B, SparseWeightVector v)
Sets the name of the classifier, the learning rate, beta and the weight vector to the specified values.

Method Summary

double	getBeta() Returns the current value of the beta member variable.
void	learn(java.lang.Object example) Evaluates the example object with the LinearThresholdUnit.score(Object), updates the learning rate, and finally updates LinearThresholdUnit.weightVector as described above.
private static boolean	nearlyEqualTo(double a, double b) Determines if a is nearly equal to b based on the value of the TOLERANCE member variable.
void	setBeta(double B) Sets the beta member variable to the specified value.
private void	updateLearningRate(java.lang.Object example, double s, boolean label) Updates the SparsePerceptron.learningRate member variable.
void	write(java.io.PrintStream out) Writes the algorithm's internal representation as text.

Methods inherited from class LBJ2.learn.SparsePerceptron

clone, demote, getLearningRate, promote, setLearningRate

Methods inherited from class LBJ2.learn.LinearThresholdUnit

allowableValues, classify, forget, getInitialWeight, getNegativeThickness, getPositiveThickness, getThreshold, score, scores, setInitialWeight, setLabeler, setNegativeThickness, setPositiveThickness, setThickness, setThreshold

Methods inherited from class LBJ2.learn.Learner

doneLearning, getExtractor, getLabeler, learn, save, setExtractor

Methods inherited from class LBJ2.classify.Classifier

binaryRead, binaryRead, binaryRead, binaryRead, binaryWrite, binaryWrite, classify, discreteValue, discreteValueArray, getCompositeChildren, getInputType, getOutputType, realValue, realValueArray, test, toString, valueIndexOf

Methods inherited from class java.lang.Object

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

Field Detail

TOLERANCE

public static final double TOLERANCE

Used to decide if two values are nearly equal to each other.

See Also:

nearlyEqualTo(double,double), Constant Field Values

defaultBeta

public static final double defaultBeta

Default value for beta.

See Also:

Constant Field Values

beta

protected double beta

The user supplied learning algorithm parameter; default defaultBeta. The learning rate changes as a function of beta.

saveLearningRate

protected double saveLearningRate

A copy of the original learning rate value is stored here so that SparsePerceptron.learningRate can be modified without losing it.

Constructor Detail

BinaryMIRA

public BinaryMIRA()

The learning rate and beta take default values while the name of the classifier takes the empty string.

BinaryMIRA

public BinaryMIRA(double r)

Sets the learning rate to the specified value, and beta to the default, while the name of the classifier takes the empty string.

Parameters:

r - The desired learning rate value.

BinaryMIRA

public BinaryMIRA(double r,
                  double B)

Sets the learning rate and beta to the specified values, while the name of the classifier takes the empty string.

Parameters:

r - The desired learning rate value.

B - the desired beta value.

BinaryMIRA

public BinaryMIRA(double r,
                  double B,
                  SparseWeightVector v)

Sets the learning rate, beta and the weight vector to the specified values.

Parameters:

r - The desired learning rate.

B - The desired beta value.

v - The desired weight vector.

BinaryMIRA

public BinaryMIRA(BinaryMIRA.Parameters p)

Initializing constructor. Sets all member variables to their associated settings in the BinaryMIRA.Parameters object.

Parameters:

p - The settings of all parameters.

BinaryMIRA

public BinaryMIRA(java.lang.String n)

Sets the name of the classifier to the specified value, while the learning rate and beta take default values.

Parameters:

n - The name of the classifier.

BinaryMIRA

public BinaryMIRA(java.lang.String n,
                  double r)

Sets the name of the classifier and learning rate to the specified values, while beta takes the default value.

Parameters:

n - The name of the classifier.

r - The desired learning rate.

BinaryMIRA

public BinaryMIRA(java.lang.String n,
                  double r,
                  double B)

Sets the name of the classifier, the learning rate and beta to the specified values.

Parameters:

n - The name of the classifier.

r - The desired learning rate.

B - The desired beta value.

BinaryMIRA

public BinaryMIRA(java.lang.String n,
                  double r,
                  double B,
                  SparseWeightVector v)

Sets the name of the classifier, the learning rate, beta and the weight vector to the specified values. Use this constructor to specify an alternative subclass of SparseWeightVector.

Parameters:

n - The name of the classifier.

r - The desired learning rate.

B - The desired beta value.

v - The desired weight vector.

BinaryMIRA

public BinaryMIRA(java.lang.String n,
                  BinaryMIRA.Parameters p)

Initializing constructor. Sets all member variables to their associated settings in the BinaryMIRA.Parameters object.

Parameters:

n - The name of the classifier.

p - The settings of all parameters.

Method Detail

getBeta

public double getBeta()

Returns the current value of the beta member variable.

Returns:

The value of the beta variable.

setBeta

public void setBeta(double B)

Sets the beta member variable to the specified value.

Parameters:

B - The new value for beta.

learn

public void learn(java.lang.Object example)

Evaluates the example object with the LinearThresholdUnit.score(Object), updates the learning rate, and finally updates LinearThresholdUnit.weightVector as described above.

Overrides:

learn in class LinearThresholdUnit

Parameters:

example - The example object

nearlyEqualTo

private static boolean nearlyEqualTo(double a,
                                     double b)

Determines if a is nearly equal to b based on the value of the TOLERANCE member variable.

Parameters:

a - The first value

b - The second value

Returns:

True if they are nearly equal, false otherwise.

updateLearningRate

private void updateLearningRate(java.lang.Object example,
                                double s,
                                boolean label)

Updates the SparsePerceptron.learningRate member variable.

Parameters:

example - The example object

s - The score for the example object

label - The label

write

public void write(java.io.PrintStream out)

Writes the algorithm's internal representation as text. In the first line of output, the name of the classifier is printed, followed by SparsePerceptron.learningRate, beta, LinearThresholdUnit.initialWeight, LinearThresholdUnit.threshold, LinearThresholdUnit.positiveThickness, LinearThresholdUnit.negativeThickness, and finally LinearThresholdUnit.bias.

Overrides:

write in class SparsePerceptron

Parameters:

out - The output stream.