/*
* GMRFSkyrideFixedEffectsGibbsOperator.java
*
* Copyright (c) 2002-2015 Alexei Drummond, Andrew Rambaut and Marc Suchard
*
* This file is part of BEAST.
* See the NOTICE file distributed with this work for additional
* information regarding copyright ownership and licensing.
*
* BEAST is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* BEAST is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with BEAST; if not, write to the
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301 USA
*/
package dr.evomodel.coalescent.operators;
import dr.evomodel.coalescent.GMRFSkyrideLikelihood;
import dr.evomodelxml.coalescent.operators.GMRFSkyrideFixedEffectsGibbsOperatorParser;
import dr.inference.model.Parameter;
import dr.inference.operators.GibbsOperator;
import dr.inference.operators.SimpleMCMCOperator;
import dr.math.distributions.MultivariateDistribution;
import no.uib.cipr.matrix.*;
/**
* A Gibbs operator to update the population size parameters under a Gaussian Markov random field prior
*
* @author Erik Bloomquist
* @author Vladimir Minin
* @author Marc Suchard
* @version $Id: GMRFSkylineFixedEffectsGibbsOperator.java,v 1.5 2007/03/20 11:26:49 msuchard Exp $
*/
public class GMRFSkyrideFixedEffectsGibbsOperator extends SimpleMCMCOperator implements GibbsOperator {
private GMRFSkyrideLikelihood gmrfLikelihood;
private DenseVector mean;
private DenseMatrix precision;
private int fieldLength;
private int dim;
public GMRFSkyrideFixedEffectsGibbsOperator(Parameter param,
GMRFSkyrideLikelihood gmrfLikelihood, MultivariateDistribution prior, double weight) {
this.gmrfLikelihood = gmrfLikelihood;
mean = new DenseVector(prior.getMean());
precision = new DenseMatrix(prior.getScaleMatrix());
this.fieldLength = gmrfLikelihood.getPopSizeParameter().getDimension();
this.dim = param.getDimension();
this.gmrfLikelihood = gmrfLikelihood;
setWeight(weight);
}
public double doOperation() {
DenseMatrix X = new DenseMatrix(gmrfLikelihood.getDesignMatrix().getParameterAsMatrix());
SymmTridiagMatrix Q = gmrfLikelihood.getScaledWeightMatrix(gmrfLikelihood.getPrecisionParameter().getParameterValue(0),
gmrfLikelihood.getLambdaParameter().getParameterValue(0));
DenseVector gamma = new DenseVector(gmrfLikelihood.getPopSizeParameter().getParameterValues());
Parameter.Abstract beta = (Parameter.Abstract) gmrfLikelihood.getBetaParameter();
//Set up the Vectors and matricies for the gibbs step
DenseMatrix gibbsPrecision = precision.copy();
UpperSPDDenseMatrix gibbsVariance;
DenseVector gibbsMean = new DenseVector(dim);
DenseMatrix workingMatrix = new DenseMatrix(dim, fieldLength);
DenseVector workingVector = new DenseVector(dim);
//Get the correct forms
X.transAmultAdd(Q, workingMatrix);
workingMatrix.multAdd(X, gibbsPrecision);
precision.mult(mean, workingVector);
workingMatrix.multAdd(gamma, workingVector);
workingMatrix = Matrices.identity(dim);
gibbsPrecision.solve(Matrices.identity(dim), workingMatrix);
gibbsVariance = new UpperSPDDenseMatrix(workingMatrix);
gibbsVariance.mult(workingVector, gibbsMean);
//Propose a new value for beta
DenseVector betaNew = GMRFSkyrideBlockUpdateOperator.getMultiNormal(gibbsMean, gibbsVariance);
for (int i = 0; i < dim; i++) {
beta.setParameterValueQuietly(i, betaNew.get(i));
}
beta.fireParameterChangedEvent();
return 0;
}
public int getStepCount() {
return 0;
}
public String getPerformanceSuggestion() {
return null;
}
//MCMCOperator INTERFACE
public final String getOperatorName() {
return GMRFSkyrideFixedEffectsGibbsOperatorParser.GMRF_GIBBS_OPERATOR;
}
}