io.github.ericmedvet.jgea.problem.regression.univariate.UnivariateRegressionFitness Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of jgea.problem Show documentation
Problem (benchmarks and templates) for jgea.
The newest version!
/*-
 * ========================LICENSE_START=================================
 * jgea-problem
 * %%
 * Copyright (C) 2018 - 2024 Eric Medvet
 * %%
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 * =========================LICENSE_END==================================
 */

package io.github.ericmedvet.jgea.problem.regression.univariate;

import io.github.ericmedvet.jgea.core.fitness.CaseBasedFitness;
import io.github.ericmedvet.jgea.core.representation.NamedUnivariateRealFunction;
import io.github.ericmedvet.jgea.problem.regression.NumericalDataset;
import java.util.List;
import java.util.Map;
import java.util.function.BiFunction;
import java.util.function.Function;
import java.util.function.IntFunction;
import java.util.stream.IntStream;

public class UnivariateRegressionFitness
    implements CaseBasedFitness, Double, Double> {

  private final NumericalDataset dataset;
  private final Metric metric;

  private List actualYs;

  public UnivariateRegressionFitness(NumericalDataset dataset, Metric metric) {
    this.dataset = dataset;
    this.metric = metric;
    actualYs = null;
  }

  public enum Metric implements Function, Double> {
    MAE(ys -> ys.stream()
        .mapToDouble(y -> Math.abs(y.predicted - y.actual))
        .average()
        .orElse(Double.NaN)),
    MSE(ys -> ys.stream()
        .mapToDouble(y -> (y.predicted - y.actual) * (y.predicted - y.actual))
        .average()
        .orElse(Double.NaN)),
    RMSE(ys -> Math.sqrt(ys.stream()
        .mapToDouble(y -> (y.predicted - y.actual) * (y.predicted - y.actual))
        .average()
        .orElse(Double.NaN))),
    NMSE(ys -> ys.stream()
            .mapToDouble(y -> (y.predicted - y.actual) * (y.predicted - y.actual))
            .average()
            .orElse(Double.NaN)
        / ys.stream().mapToDouble(y -> y.actual).average().orElse(1d));
    private final Function, Double> function;

    Metric(Function, Double> function) {
      this.function = function;
    }

    @Override
    public Double apply(List ys) {
      return function.apply(ys);
    }
  }

  private record Y(double predicted, double actual) {}

  public static List pairs(List predictedYs, List actualYs) {
    return IntStream.range(0, actualYs.size())
        .mapToObj(i -> new Y(predictedYs.get(i), actualYs.get(i)))
        .toList();
  }

  public NumericalDataset getDataset() {
    return dataset;
  }

  public Metric getMetric() {
    return metric;
  }

  @Override
  public Function, Double> aggregateFunction() {
    return predictedYs -> {
      if (actualYs == null) {
        actualYs = IntStream.range(0, dataset.size())
            .mapToObj(i -> dataset.exampleProvider().apply(i).ys()[0])
            .toList();
      }
      return metric.apply(pairs(predictedYs, actualYs));
    };
  }

  @Override
  public BiFunction, Double> caseFunction() {
    return NamedUnivariateRealFunction::computeAsDouble;
  }

  @Override
  public IntFunction> caseProvider() {
    return i -> dataset.namedExampleProvider().apply(i).x();
  }

  @Override
  public int nOfCases() {
    return dataset.size();
  }
}