ContactCenters
V. 0.9.9.

umontreal.iro.lecuyer.contactcenters.contact
Class CorrelationMtxFitting

java.lang.Object
  extended by umontreal.iro.lecuyer.contactcenters.contact.CorrelationMtxFitting

public class CorrelationMtxFitting
extends Object

Fits the parametric model for the correlation matrix using method of least squares. Two models are implemented: (i) general linear model rj = abj + c, (ii) single rho Markov model rj = bj. Method fitMarkovGeneralLinear() implements fitting of model (i), method fitMarkovSingleRho() implements fitting of model (ii). The optimization of parameter b is performed using exhaustive grid search with step step in the range [-1+delta, 1-delta]. For model (i) parameters a and c have closed form expressions in terms of the entries of correlation matrix and parameter b.

Author:
Boris N. Oreshkin

Constructor Summary
CorrelationMtxFitting(double[] CorrVector)
           
CorrelationMtxFitting(double[] CorrVector, double tol)
           
 
Method Summary
 double[] fitMarkovGeneralLinear()
          Fits general linear model rj = abj + c.
 double fitMarkovSingleRho()
          Fits the Markov model with single correlation coefficient of the form ρj = bj.
 void setDelta(double delta)
          Sets the limits in which exhaustive grid search for the optimization of parameter b is performed.
 void setStep(double step)
          Sets the exhaustive grid search grid size for the optimization of parameter b.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

CorrelationMtxFitting

public CorrelationMtxFitting(double[] CorrVector)

CorrelationMtxFitting

public CorrelationMtxFitting(double[] CorrVector,
                             double tol)
Method Detail

setDelta

public void setDelta(double delta)
Sets the limits in which exhaustive grid search for the optimization of parameter b is performed. The search is performed in the interval [-1+delta, 1-delta].

Parameters:
delta - Exhaustive grid search boundary offset

setStep

public void setStep(double step)
Sets the exhaustive grid search grid size for the optimization of parameter b.

Parameters:
step - Exhaustive grid search grid size

fitMarkovGeneralLinear

public double[] fitMarkovGeneralLinear()
Fits general linear model rj = abj + c. Returns vector of length 3 with parameters a, b and c.

Returns:
vector of [a, b and c].

fitMarkovSingleRho

public double fitMarkovSingleRho()
Fits the Markov model with single correlation coefficient of the form ρj = bj.

Returns:
parameter b.

ContactCenters
V. 0.9.9.

To submit a bug or ask questions, send an e-mail to Richard Simard.