package jsat.distributions.multivariate;
import java.util.List;
import java.util.concurrent.ExecutorService;
import jsat.DataSet;
import jsat.classifiers.DataPoint;
import jsat.linear.DenseVector;
import jsat.linear.Vec;
/**
* Common class for implementing a multivariate distribution. A number of methods are pre implemented,
* building off of the implementation of the remaining methods.
* Note: the default implementation for the multithreaded methods calls the non threaded version of the method.
* The exception to this is the {@link #setUsingData(jsat.DataSet, java.util.concurrent.ExecutorService) } method,
* which calls {@link #setUsingData(java.util.List, java.util.concurrent.ExecutorService) }
*
* @author Edward Raff
*/
public abstract class MultivariateDistributionSkeleton implements MultivariateDistribution
{
private static final long serialVersionUID = 4080753806798149915L;
@Override
public double logPdf(Vec x)
{
double logPDF = Math.log(pdf(x));
if(Double.isInfinite(logPDF) && logPDF < 0)//log(0) == -Infinty
return -Double.MAX_VALUE;
return logPDF;
}
@Override
abstract public MultivariateDistribution clone();
}
Ads help maintain this website.