package jsat.distributions.kernels;
import java.util.List;
import jsat.linear.Vec;
import jsat.linear.distancemetrics.DistanceMetric;
import jsat.parameters.Parameter;
import jsat.parameters.Parameter.ParameterHolder;
/**
* This abstract class provides the means of implementing a Kernel based off
* some {@link DistanceMetric}. This will pre-implement most of the methods of
* the KernelTrick interface, including using the distance acceleration of the
* metric (if supported) when appropriate.
*
* @author Edward Raff
*/
public abstract class DistanceMetricBasedKernel implements KernelTrick
{
private static final long serialVersionUID = 8395066824809874527L;
/**
* the distance metric to use for the Kernel
*/
@ParameterHolder
protected DistanceMetric d;
/**
* Creates a new distance based kerenel
* @param d the distance metric to use
*/
public DistanceMetricBasedKernel(DistanceMetric d)
{
this.d = d;
}
@Override
abstract public KernelTrick clone();
@Override
public boolean supportsAcceleration()
{
return d.supportsAcceleration();
}
@Override
public List getAccelerationCache(List extends Vec> trainingSet)
{
return d.getAccelerationCache(trainingSet);
}
@Override
public List getQueryInfo(Vec q)
{
return d.getQueryInfo(q);
}
@Override
public void addToCache(Vec newVec, List cache)
{
cache.addAll(d.getQueryInfo(newVec));
}
@Override
public double evalSum(List extends Vec> finalSet, List cache, double[] alpha, Vec y, int start, int end)
{
return evalSum(finalSet, cache, alpha, y, d.getQueryInfo(y), start, end);
}
@Override
public double evalSum(List extends Vec> finalSet, List cache, double[] alpha, Vec y, List qi, int start, int end)
{
double sum = 0;
for (int i = start; i < end; i++)
if (alpha[i] != 0)
sum += alpha[i] * eval(i, y, qi, finalSet, cache);
return sum;
}
}
Ads help maintain this website.