[11683] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[17181] | 3 | * Copyright (C) Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[11683] | 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 HeuristicLab.Common;
|
---|
| 25 |
|
---|
| 26 | namespace HeuristicLab.Problems.DataAnalysis {
|
---|
[14801] | 27 | public class ClassificationPerformanceMeasuresCalculator : DeepCloneable {
|
---|
[11683] | 28 |
|
---|
[11685] | 29 | public ClassificationPerformanceMeasuresCalculator(string positiveClassName, double positiveClassValue) {
|
---|
| 30 | this.positiveClassName = positiveClassName;
|
---|
[11683] | 31 | this.positiveClassValue = positiveClassValue;
|
---|
[11761] | 32 | Reset();
|
---|
[11683] | 33 | }
|
---|
| 34 |
|
---|
[14801] | 35 | protected ClassificationPerformanceMeasuresCalculator(ClassificationPerformanceMeasuresCalculator original, Cloner cloner)
|
---|
| 36 | : base(original, cloner) {
|
---|
| 37 | positiveClassName = original.positiveClassName;
|
---|
| 38 | positiveClassValue = original.positiveClassValue;
|
---|
| 39 | truePositiveCount = original.truePositiveCount;
|
---|
| 40 | falsePositiveCount = original.falsePositiveCount;
|
---|
| 41 | trueNegativeCount = original.trueNegativeCount;
|
---|
| 42 | falseNegativeCount = original.falseNegativeCount;
|
---|
| 43 | errorState = original.errorState;
|
---|
| 44 | }
|
---|
| 45 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
| 46 | return new ClassificationPerformanceMeasuresCalculator(this, cloner);
|
---|
| 47 | }
|
---|
| 48 |
|
---|
[11683] | 49 | #region Properties
|
---|
| 50 | private int truePositiveCount, falsePositiveCount, trueNegativeCount, falseNegativeCount;
|
---|
| 51 |
|
---|
[11761] | 52 | private readonly string positiveClassName;
|
---|
[11685] | 53 | public string PositiveClassName {
|
---|
[11761] | 54 | get { return positiveClassName; }
|
---|
[11685] | 55 | }
|
---|
[11761] | 56 |
|
---|
| 57 | private readonly double positiveClassValue;
|
---|
[11685] | 58 | public double PositiveClassValue {
|
---|
[11761] | 59 | get { return positiveClassValue; }
|
---|
[11685] | 60 | }
|
---|
[11683] | 61 | public double TruePositiveRate {
|
---|
| 62 | get {
|
---|
| 63 | double divisor = truePositiveCount + falseNegativeCount;
|
---|
| 64 | return divisor.IsAlmost(0) ? double.NaN : truePositiveCount / divisor;
|
---|
| 65 | }
|
---|
| 66 | }
|
---|
| 67 | public double TrueNegativeRate {
|
---|
| 68 | get {
|
---|
| 69 | double divisor = falsePositiveCount + trueNegativeCount;
|
---|
| 70 | return divisor.IsAlmost(0) ? double.NaN : trueNegativeCount / divisor;
|
---|
| 71 | }
|
---|
| 72 | }
|
---|
| 73 | public double PositivePredictiveValue {
|
---|
| 74 | get {
|
---|
| 75 | double divisor = truePositiveCount + falsePositiveCount;
|
---|
| 76 | return divisor.IsAlmost(0) ? double.NaN : truePositiveCount / divisor;
|
---|
| 77 | }
|
---|
| 78 | }
|
---|
| 79 | public double NegativePredictiveValue {
|
---|
| 80 | get {
|
---|
| 81 | double divisor = trueNegativeCount + falseNegativeCount;
|
---|
| 82 | return divisor.IsAlmost(0) ? double.NaN : trueNegativeCount / divisor;
|
---|
| 83 | }
|
---|
| 84 | }
|
---|
| 85 | public double FalsePositiveRate {
|
---|
| 86 | get {
|
---|
| 87 | double divisor = falsePositiveCount + trueNegativeCount;
|
---|
| 88 | return divisor.IsAlmost(0) ? double.NaN : falsePositiveCount / divisor;
|
---|
| 89 | }
|
---|
| 90 | }
|
---|
| 91 | public double FalseDiscoveryRate {
|
---|
| 92 | get {
|
---|
| 93 | double divisor = falsePositiveCount + truePositiveCount;
|
---|
| 94 | return divisor.IsAlmost(0) ? double.NaN : falsePositiveCount / divisor;
|
---|
| 95 | }
|
---|
| 96 | }
|
---|
| 97 |
|
---|
| 98 | private OnlineCalculatorError errorState;
|
---|
| 99 | public OnlineCalculatorError ErrorState {
|
---|
| 100 | get { return errorState; }
|
---|
| 101 | }
|
---|
| 102 | #endregion
|
---|
[11761] | 103 |
|
---|
[11683] | 104 | public void Reset() {
|
---|
| 105 | truePositiveCount = 0;
|
---|
| 106 | falseNegativeCount = 0;
|
---|
| 107 | trueNegativeCount = 0;
|
---|
| 108 | falseNegativeCount = 0;
|
---|
| 109 | errorState = OnlineCalculatorError.InsufficientElementsAdded;
|
---|
| 110 | }
|
---|
| 111 |
|
---|
| 112 | public void Add(double originalClassValue, double estimatedClassValue) {
|
---|
| 113 | // ignore cases where original is NaN completely
|
---|
[11761] | 114 | if (double.IsNaN(originalClassValue)) return;
|
---|
| 115 |
|
---|
| 116 | if (originalClassValue.IsAlmost(positiveClassValue)
|
---|
| 117 | || estimatedClassValue.IsAlmost(positiveClassValue)) { //positive class/positive class estimation
|
---|
| 118 | if (estimatedClassValue.IsAlmost(originalClassValue)) {
|
---|
| 119 | truePositiveCount++;
|
---|
| 120 | } else {
|
---|
| 121 | if (estimatedClassValue.IsAlmost(positiveClassValue)) //misclassification of the negative class
|
---|
| 122 | falsePositiveCount++;
|
---|
| 123 | else //misclassification of the positive class
|
---|
| 124 | falseNegativeCount++;
|
---|
[11683] | 125 | }
|
---|
[11761] | 126 | } else { //negative class/negative class estimation
|
---|
| 127 | //In a multiclass classification all misclassifications of the negative class
|
---|
| 128 | //will be treated as true negatives except on positive class estimations
|
---|
| 129 | trueNegativeCount++;
|
---|
[11683] | 130 | }
|
---|
[11761] | 131 |
|
---|
| 132 | errorState = OnlineCalculatorError.None; // number of (non-NaN) samples >= 1
|
---|
[11683] | 133 | }
|
---|
| 134 |
|
---|
[11761] | 135 | public void Calculate(IEnumerable<double> originalClassValues, IEnumerable<double> estimatedClassValues) {
|
---|
[11683] | 136 | IEnumerator<double> originalEnumerator = originalClassValues.GetEnumerator();
|
---|
| 137 | IEnumerator<double> estimatedEnumerator = estimatedClassValues.GetEnumerator();
|
---|
| 138 |
|
---|
| 139 | // always move forward both enumerators (do not use short-circuit evaluation!)
|
---|
| 140 | while (originalEnumerator.MoveNext() & estimatedEnumerator.MoveNext()) {
|
---|
| 141 | double original = originalEnumerator.Current;
|
---|
| 142 | double estimated = estimatedEnumerator.Current;
|
---|
| 143 | Add(original, estimated);
|
---|
| 144 | if (ErrorState != OnlineCalculatorError.None) break;
|
---|
| 145 | }
|
---|
| 146 |
|
---|
| 147 | // check if both enumerators are at the end to make sure both enumerations have the same length
|
---|
[11761] | 148 | if (ErrorState == OnlineCalculatorError.None && (estimatedEnumerator.MoveNext() || originalEnumerator.MoveNext())) {
|
---|
[11683] | 149 | throw new ArgumentException("Number of elements in originalValues and estimatedValues enumerations doesn't match.");
|
---|
| 150 | }
|
---|
[11761] | 151 | errorState = ErrorState;
|
---|
[11683] | 152 | }
|
---|
| 153 | }
|
---|
| 154 | }
|
---|