1  #region License Information


2  /* HeuristicLab


3  * Copyright (C) 20022012 Heuristic and Evolutionary Algorithms Laboratory (HEAL)


4  *


5  * This file is part of HeuristicLab.


6  *


7  * HeuristicLab is free software: you can redistribute it and/or modify


8  * it under the terms of the GNU General Public License as published by


9  * the Free Software Foundation, either version 3 of the License, or


10  * (at your option) any later version.


11  *


12  * HeuristicLab is distributed in the hope that it will be useful,


13  * but WITHOUT ANY WARRANTY; without even the implied warranty of


14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the


15  * GNU General Public License for more details.


16  *


17  * You should have received a copy of the GNU General Public License


18  * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.


19  */


20  #endregion


21 


22  using System;


23  using System.Collections.Generic;


24  using System.Linq;


25  using HeuristicLab.Common;


26  using HeuristicLab.Core;


27  using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;


28 


29  namespace HeuristicLab.Problems.DataAnalysis {


30  /// <summary>


31  ///


32  /// </summary>


33  [StorableClass]


34  [Item("ContinuousPointCertaintyWeightCalculator", "")]


35  public class ContinuousPointCertaintyWeightCalculator : DiscriminantClassificationWeightCalculator {


36 


37  public ContinuousPointCertaintyWeightCalculator()


38  : base() {


39  }


40 


41  [StorableConstructor]


42  protected ContinuousPointCertaintyWeightCalculator(bool deserializing) : base(deserializing) { }


43  protected ContinuousPointCertaintyWeightCalculator(ContinuousPointCertaintyWeightCalculator original, Cloner cloner)


44  : base(original, cloner) {


45  }


46 


47  public override IDeepCloneable Clone(Cloner cloner) {


48  return new ContinuousPointCertaintyWeightCalculator(this, cloner);


49  }


50 


51  protected override IEnumerable<double> DiscriminantCalculateWeights(IEnumerable<IDiscriminantFunctionClassificationSolution> discriminantSolutions) {


52  List<double> weights = new List<double>();


53  IClassificationProblemData problemData = discriminantSolutions.ElementAt(0).ProblemData;


54  IEnumerable<double> targetValues;


55  IEnumerator<double> trainingValues;


56 


57  //only works for binary classification


58  if (!problemData.ClassValues.Count().Equals(2))


59  return Enumerable.Repeat<double>(1, discriminantSolutions.Count());


60 


61  double maxClass = problemData.ClassValues.Max();


62  double minClass = problemData.ClassValues.Min();


63  double halfDistanceBetweenClasses = (maxClass  minClass) / 2;


64 


65  foreach (var solution in discriminantSolutions) {


66  problemData = solution.ProblemData;


67  targetValues = GetValues(problemData.Dataset.GetDoubleValues(problemData.TargetVariable).ToList(), problemData.TrainingIndizes);


68  trainingValues = targetValues.GetEnumerator();


69 


70  IEnumerator<double> estimatedTrainingVal = solution.EstimatedTrainingValues.GetEnumerator();


71  IEnumerator<double> estimatedTrainingClassVal = solution.EstimatedTrainingClassValues.GetEnumerator();


72 


73  double curWeight = 0.0;


74  while (estimatedTrainingVal.MoveNext() && estimatedTrainingClassVal.MoveNext() && trainingValues.MoveNext()) {


75  if (trainingValues.Current.Equals(maxClass)) {


76  if (estimatedTrainingVal.Current >= maxClass)


77  curWeight += 1.0;


78  else {


79  double distanceToPoint = Math.Abs(estimatedTrainingVal.Current  maxClass);


80  curWeight += Math.Max(1.0  (1.0 / halfDistanceBetweenClasses) * distanceToPoint, 1.0);


81  }


82  } else if (trainingValues.Current.Equals(minClass)) {


83  if (estimatedTrainingVal.Current <= minClass)


84  curWeight += 1.0;


85  else {


86  double distanceToPoint = Math.Abs(estimatedTrainingVal.Current  minClass);


87  curWeight += Math.Max(1.0  (1.0 / halfDistanceBetweenClasses) * distanceToPoint, 1.0);


88  }


89  }


90  }


91  // normalize the weight (otherwise a model with a bigger training partition would probably be better)


92  weights.Add(curWeight / targetValues.Count());


93  }


94  return weights;


95  }


96  }


97  }

