epfl.pdf
Class PEMOnMultNormalKernel

java.lang.Object
  |
  +--epfl.pdf.EMOnMultNormalKernel
        |
        +--epfl.pdf.PEMOnMultNormalKernel

public class PEMOnMultNormalKernel
extends EMOnMultNormalKernel

EM algorithm on multivariate normal kernel pdf with prior probabilities


Fields inherited from class epfl.pdf.EMOnMultNormalKernel
minGrowth
 
Constructor Summary
PEMOnMultNormalKernel(MultNormalKernelPdf workingPdf, double minGrowth, double minSigma, VectorReader data)
          Creates a new EM algorithm on the underlying NormalKernel pdf
PEMOnMultNormalKernel(MultNormalKernelPdf workingPdf, double minGrowth, double minSigma, VectorReader data, VectorWriter likehoodOut)
          Creates a new EM algorithm on the underlying MultNormalKernel pdf
 
Methods inherited from class epfl.pdf.EMOnMultNormalKernel
close, currentGrowth, currentLikehood, currentStep, flush, getNormalKernelPdf, init, maximizeLikehood, nextStep
 
Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

PEMOnMultNormalKernel

public PEMOnMultNormalKernel(MultNormalKernelPdf workingPdf,
                             double minGrowth,
                             double minSigma,
                             VectorReader data,
                             VectorWriter likehoodOut)
Creates a new EM algorithm on the underlying MultNormalKernel pdf
Parameters:
workingPdf - the underlying MultNormalKernel pdf
minGrowth - convergence criteria
minSigma - min value for sigma
data - the data for building a pdf
data dimension = working pdf dimension + 1 for the prior probability
likehoodOut - the output stream for likehood maximization

PEMOnMultNormalKernel

public PEMOnMultNormalKernel(MultNormalKernelPdf workingPdf,
                             double minGrowth,
                             double minSigma,
                             VectorReader data)
Creates a new EM algorithm on the underlying NormalKernel pdf
Parameters:
workingPdf - the underlying NormalKernel pdf
minGrowth - convergence criteria
minSigma - min value for sigma
data - the data for building a pdf
data dimension = working pdf dimension + 1 for the prior probability