com.expleague.ml.loss.blockwise.BlockwiseAccuracyLogit 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.Mx;
import com.expleague.commons.math.vectors.Vec;
import com.expleague.commons.seq.IntSeq;
import com.expleague.ml.data.set.DataSet;

import static java.lang.Math.exp;

public class BlockwiseAccuracyLogit extends BlockwiseMLLLogit {
  public BlockwiseAccuracyLogit(IntSeq target, DataSet owner) {
    super(target, owner);
  }

  public BlockwiseAccuracyLogit(Vec target, DataSet owner) {
    super(target, owner);
  }

  @Override
  public double value(Vec pointBlock, int blockId) {
    int result = blockSize();
    double score = 1;
    for (int c = 0; c < blockSize(); c++) {
      double current = exp(pointBlock.get(c));
      if (current > score) {
        result = c;
        score = current;
      }
    }
    final int pointClass = label(blockId);
    return pointClass == result ? 1 : 0;
  }

  @Override
  public double transformResultValue(double value) {
    return value / labels().length();
  }
}