com.expleague.ml.loss.blockwise.BlockwiseWeightedLoss Maven / Gradle / Ivy

Go to download

Show more of this group Show more artifacts with this name
Show all versions of jmll

Various ML methods implemented by myself and my students

There is a newer version: 1.4.9

package com.expleague.ml.loss.blockwise;

import com.expleague.commons.math.vectors.Vec;
import com.expleague.commons.math.vectors.VecTools;
import com.expleague.ml.BlockwiseFuncC1;
import com.expleague.ml.TargetFunc;
import com.expleague.ml.data.set.DataSet;

/**
* User: qdeee
* Date: 26.11.13
* Time: 9:54
*/

public class BlockwiseWeightedLoss extends BlockwiseFuncC1.Stub implements TargetFunc{
  private final BasedOn metric;
  private final int[] weights;

  public BlockwiseWeightedLoss(final BasedOn metric, final int[] weights) {
    if (metric.dim() / metric.blockSize() != weights.length)
      throw new IllegalArgumentException("weights.length must be equal to blocks count");
    this.metric = metric;
    this.weights = weights;
  }

  @Override
  public int dim() {
    return metric.xdim();
  }

  public double weight(final int index) {
    return weights[index];
  }

  public BasedOn base() {
    return metric;
  }

  @Override
  public void gradient(final Vec pointBlock, final Vec result, final int index) {
    if (weights[index] > 0) {
      metric.gradient(pointBlock, result, index);
      VecTools.scale(result, weights[index]);
    }
  }

  @Override
  public double value(final Vec pointBlock, final int index) {
    return weights[index] > 0 ? weights[index] * metric.value(pointBlock, index) : 0;
  }

  @Override
  public double transformResultValue(final double value) {
    return metric.transformResultValue(value);
  }

  @Override
  public int blockSize() {
    return metric.blockSize();
  }

  @Override
  public DataSet owner() {
    return metric.owner();
  }
}