diva.sketch.classification

Class GaussianWeightSet

java.lang.Object

diva.sketch.classification.WeightSet

diva.sketch.classification.GaussianWeightSet

public class GaussianWeightSet
extends WeightSet

Given a set of training examples (each example is a feature vector), a Gaussian classifier computes the mu and sigma for each type of features. The mu of a particular type of feature is calculated by taking the mean of the features of that type from the examples. The sigma is the standard deviation.

Version:

$Revision: 1.4 $

Author:

Heloise Hse (hwawen@eecs.berkeley.edu), Michael Shilman (michaels@eecs.berkeley.edu)

Constructor Summary

Constructors

Constructor and Description
GaussianWeightSet(String type) Constructs a gaussian weight set with the specified type.

Method Summary

Modifier and Type	Method and Description
protected FeatureSet	computeMu() Calculates the mu value for Gaussian distribution by taking the mean of each feature type.
protected FeatureSet	computeSigma() Calculates the sigma of each feature in this classifier.

Methods inherited from class diva.sketch.classification.WeightSet

addExample, clearExamples, debug, examples, getExampleCount, getMuValues, getSigmaValues, getType, toString, train

Methods inherited from class java.lang.Object

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

Constructor Detail

GaussianWeightSet

public GaussianWeightSet(String type)

Constructs a gaussian weight set with the specified type.

Method Detail

computeMu

protected FeatureSet computeMu()
                        throws ClassifierException

Calculates the mu value for Gaussian distribution by taking the mean of each feature type.

Specified by:

computeMu in class WeightSet

Throws:

ClassifierException

computeSigma

protected FeatureSet computeSigma()
                           throws ClassifierException

Calculates the sigma of each feature in this classifier. In Gaussian distribution, sigma equals standard deviation. This method assumes that the mu values have already been calculated, and will throw an exception if this is not the case.

   sigma[i] = (sqrt (sum (X[i]-u)^2)))/N
 

Specified by:

computeSigma in class WeightSet

Throws:

ClassifierException