com.expleague.ml.loss.multiclass.MCMicroPrecision 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.multiclass;

import com.expleague.commons.math.vectors.Vec;
import com.expleague.commons.math.vectors.VecIterator;
import com.expleague.commons.seq.IntSeq;
import com.expleague.commons.math.Func;
import com.expleague.ml.data.set.DataSet;

/**
 * User: qdeee
 * Date: 09.04.14
 */
public class MCMicroPrecision extends Func.Stub implements ClassicMulticlassLoss {
  protected final IntSeq target;
  private final DataSet owner;

  public MCMicroPrecision(final IntSeq target, final DataSet owner) {
    this.target = target;
    this.owner = owner;
  }

  public MCMicroPrecision(final Vec target, final DataSet owner) {
    final int[] intTarget = new int[target.length()];
    final VecIterator iter = target.nonZeroes();
    while (iter.advance()) {
      intTarget[iter.index()] = (int) iter.value();
    }
    this.target = new IntSeq(intTarget);
    this.owner = owner;
  }

  @Override
  public double value(final Vec x) {
    int tp = 0;
    int fp = 0;
    for (int i = 0; i < x.dim(); i++) {
      final int expected = target.at(i);
      final int actual = (int) x.get(i);

      //skip unrecognized class
      if (actual == -1)
        continue;

      if (actual == expected)
        tp += 1;
      else
        fp += 1;
    }
    return tp / (tp + fp + 0.);
  }

  @Override
  public int dim() {
    return target.length();
  }

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

  @Override
  public IntSeq labels() {
    return target;
  }
}