#region License Information
/* HeuristicLab
* Copyright (C) 2002-2016 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
*
* This file is part of HeuristicLab.
*
* HeuristicLab is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* HeuristicLab is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with HeuristicLab. If not, see .
*/
#endregion
using System;
using System.Collections.Generic;
using System.Linq;
using HeuristicLab.Analysis;
using HeuristicLab.Common;
using HeuristicLab.Core;
using HeuristicLab.Data;
using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
using HeuristicLab.Optimization;
using HeuristicLab.Parameters;
using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression {
[Item("SymbolicRegressionSingleObjectiveOSGAEvaluator", "An evaluator which tries to predict when a child will not be able to fullfil offspring selection criteria, to save evaluation time.")]
[StorableClass]
public class SymbolicRegressionSingleObjectiveOsgaEvaluator : SymbolicRegressionSingleObjectiveEvaluator {
private const string RelativeParentChildQualityThresholdParameterName = "RelativeParentChildQualityThreshold";
private const string RelativeFitnessEvaluationIntervalSizeParameterName = "RelativeFitnessEvaluationIntervalSize";
private const string ResultCollectionParameterName = "Results";
#region parameters
public ILookupParameter ResultCollectionParameter {
get { return (ILookupParameter)Parameters[ResultCollectionParameterName]; }
}
public IFixedValueParameter CorrectlyRejectedParameter {
get { return (IFixedValueParameter)Parameters["CorrectlyRejected"]; }
}
public IFixedValueParameter IncorrectlyRejectedParameter {
get { return (IFixedValueParameter)Parameters["IncorrectlyRejected"]; }
}
public IFixedValueParameter CorrectlyNotRejectedParameter {
get { return (IFixedValueParameter)Parameters["CorrectlyNotRejected"]; }
}
public IFixedValueParameter IncorrectlyNotRejectedParameter {
get { return (IFixedValueParameter)Parameters["IncorrectlyNotRejected"]; }
}
public IValueLookupParameter ComparisonFactorParameter {
get { return (ValueLookupParameter)Parameters["ComparisonFactor"]; }
}
public IFixedValueParameter RelativeParentChildQualityThresholdParameter {
get { return (IFixedValueParameter)Parameters[RelativeParentChildQualityThresholdParameterName]; }
}
public IFixedValueParameter RelativeFitnessEvaluationIntervalSizeParameter {
get { return (IFixedValueParameter)Parameters[RelativeFitnessEvaluationIntervalSizeParameterName]; }
}
public IScopeTreeLookupParameter ParentQualitiesParameter { get { return (IScopeTreeLookupParameter)Parameters["ParentQualities"]; } }
#endregion
#region parameter properties
public double RelativeParentChildQualityThreshold {
get { return RelativeParentChildQualityThresholdParameter.Value.Value; }
set { RelativeParentChildQualityThresholdParameter.Value.Value = value; }
}
public double RelativeFitnessEvaluationIntervalSize {
get { return RelativeFitnessEvaluationIntervalSizeParameter.Value.Value; }
set { RelativeFitnessEvaluationIntervalSizeParameter.Value.Value = value; }
}
public int CorrectlyRejected {
get { return CorrectlyRejectedParameter.Value.Value; }
set { CorrectlyRejectedParameter.Value.Value = value; }
}
public int CorrectlyNotRejected {
get { return CorrectlyNotRejectedParameter.Value.Value; }
set { CorrectlyNotRejectedParameter.Value.Value = value; }
}
public int IncorrectlyRejected {
get { return IncorrectlyRejectedParameter.Value.Value; }
set { IncorrectlyRejectedParameter.Value.Value = value; }
}
public int IncorrectlyNotRejected {
get { return IncorrectlyNotRejectedParameter.Value.Value; }
set { IncorrectlyNotRejectedParameter.Value.Value = value; }
}
#endregion
public override bool Maximization {
get { return true; }
}
public SymbolicRegressionSingleObjectiveOsgaEvaluator() {
Parameters.Add(new ValueLookupParameter("ComparisonFactor", "Determines if the quality should be compared to the better parent (1.0), to the worse (0.0) or to any linearly interpolated value between them."));
Parameters.Add(new FixedValueParameter(RelativeParentChildQualityThresholdParameterName, new PercentValue(0.1)));
Parameters.Add(new FixedValueParameter(RelativeFitnessEvaluationIntervalSizeParameterName, new PercentValue(0.1)));
Parameters.Add(new FixedValueParameter("CorrectlyRejected", new IntValue(0)));
Parameters.Add(new FixedValueParameter("IncorrectlyRejected", new IntValue(0)));
Parameters.Add(new FixedValueParameter("CorrectlyNotRejected", new IntValue(0)));
Parameters.Add(new FixedValueParameter("IncorrectlyNotRejected", new IntValue(0)));
Parameters.Add(new LookupParameter(ResultCollectionParameterName));
Parameters.Add(new ScopeTreeLookupParameter("ParentQualities") { ActualName = "Quality" });
}
[StorableHook(HookType.AfterDeserialization)]
private void AfterDeserialization() {
if (!Parameters.ContainsKey(ResultCollectionParameterName))
Parameters.Add(new LookupParameter(ResultCollectionParameterName));
if (!Parameters.ContainsKey("ParentQualities"))
Parameters.Add(new ScopeTreeLookupParameter("ParentQualities") { ActualName = "Quality" });
}
[StorableConstructor]
protected SymbolicRegressionSingleObjectiveOsgaEvaluator(bool deserializing) : base(deserializing) { }
protected SymbolicRegressionSingleObjectiveOsgaEvaluator(SymbolicRegressionSingleObjectiveOsgaEvaluator original, Cloner cloner) : base(original, cloner) { }
public override IDeepCloneable Clone(Cloner cloner) {
return new SymbolicRegressionSingleObjectiveOsgaEvaluator(this, cloner);
}
public override void ClearState() {
base.ClearState();
CorrectlyNotRejected = 0;
CorrectlyRejected = 0;
IncorrectlyNotRejected = 0;
IncorrectlyRejected = 0;
}
public override IOperation InstrumentedApply() {
var solution = SymbolicExpressionTreeParameter.ActualValue;
IEnumerable rows = GenerateRowsToEvaluate();
var interpreter = SymbolicDataAnalysisTreeInterpreterParameter.ActualValue;
var estimationLimits = EstimationLimitsParameter.ActualValue;
var problemData = ProblemDataParameter.ActualValue;
var applyLinearScaling = ApplyLinearScalingParameter.ActualValue.Value;
double quality;
var parentQualities = ParentQualitiesParameter.ActualValue;
// parent subscopes are not present during evaluation of the initial population
if (parentQualities.Length > 0) {
quality = Calculate(interpreter, solution, estimationLimits, problemData, rows, applyLinearScaling);
} else {
quality = Calculate(interpreter, solution, estimationLimits.Lower, estimationLimits.Upper, problemData, rows, applyLinearScaling);
}
QualityParameter.ActualValue = new DoubleValue(quality);
return base.InstrumentedApply();
}
public static double Calculate(ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, ISymbolicExpressionTree solution, double lowerEstimationLimit, double upperEstimationLimit, IRegressionProblemData problemData, IEnumerable rows, bool applyLinearScaling) {
IEnumerable estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, problemData.Dataset, rows);
IEnumerable targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, rows);
OnlineCalculatorError errorState;
double r;
if (applyLinearScaling) {
var rCalculator = new OnlinePearsonsRCalculator();
CalculateWithScaling(targetValues, estimatedValues, lowerEstimationLimit, upperEstimationLimit, rCalculator, problemData.Dataset.Rows);
errorState = rCalculator.ErrorState;
r = rCalculator.R;
} else {
IEnumerable boundedEstimatedValues = estimatedValues.LimitToRange(lowerEstimationLimit, upperEstimationLimit);
r = OnlinePearsonsRCalculator.Calculate(targetValues, boundedEstimatedValues, out errorState);
}
if (errorState != OnlineCalculatorError.None) return double.NaN;
return r * r;
}
private double Calculate(ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, ISymbolicExpressionTree solution, DoubleLimit estimationLimits, IRegressionProblemData problemData, IEnumerable rows, bool applyLinearScaling) {
var estimatedValues = interpreter.GetSymbolicExpressionTreeValues(solution, problemData.Dataset, rows).ToList();
var targetValues = problemData.Dataset.GetDoubleValues(problemData.TargetVariable, rows);
IEnumerator targetValuesEnumerator;
double alpha = 0, beta = 1;
if (applyLinearScaling) {
var linearScalingCalculator = new OnlineLinearScalingParameterCalculator();
targetValuesEnumerator = targetValues.GetEnumerator();
var estimatedValuesEnumerator = estimatedValues.GetEnumerator();
while (targetValuesEnumerator.MoveNext() & estimatedValuesEnumerator.MoveNext()) {
double target = targetValuesEnumerator.Current;
double estimated = estimatedValuesEnumerator.Current;
if (!double.IsNaN(estimated) && !double.IsInfinity(estimated))
linearScalingCalculator.Add(estimated, target);
}
if (linearScalingCalculator.ErrorState == OnlineCalculatorError.None && (targetValuesEnumerator.MoveNext() || estimatedValuesEnumerator.MoveNext()))
throw new ArgumentException("Number of elements in target and estimated values enumeration do not match.");
alpha = linearScalingCalculator.Alpha;
beta = linearScalingCalculator.Beta;
if (linearScalingCalculator.ErrorState != OnlineCalculatorError.None) {
alpha = 0.0;
beta = 1.0;
}
}
var scaledEstimatedValuesEnumerator = estimatedValues.Select(x => x * beta + alpha).LimitToRange(estimationLimits.Lower, estimationLimits.Upper).GetEnumerator();
targetValuesEnumerator = targetValues.GetEnumerator();
var pearsonRCalculator = new OnlinePearsonsRCalculator();
var interval = (int)Math.Floor(problemData.TrainingPartition.Size * RelativeFitnessEvaluationIntervalSize);
var i = 0;
var qualityPerInterval = new List();
while (targetValuesEnumerator.MoveNext() && scaledEstimatedValuesEnumerator.MoveNext()) {
pearsonRCalculator.Add(targetValuesEnumerator.Current, scaledEstimatedValuesEnumerator.Current);
++i;
if (i % interval == 0) {
var q = pearsonRCalculator.ErrorState != OnlineCalculatorError.None ? double.NaN : pearsonRCalculator.R;
qualityPerInterval.Add(q * q);
}
}
var r = pearsonRCalculator.ErrorState != OnlineCalculatorError.None ? double.NaN : pearsonRCalculator.R;
var actualQuality = r * r;
var parentQualities = ParentQualitiesParameter.ActualValue.Select(x => x.Value);
var minQuality = parentQualities.Min();
var maxQuality = parentQualities.Max();
var comparisonFactor = ComparisonFactorParameter.ActualValue.Value;
var parentQuality = minQuality + (maxQuality - minQuality) * comparisonFactor;
var threshold = parentQuality * RelativeParentChildQualityThreshold;
//var predictedRejected = qualityPerInterval.Any(x => double.IsNaN(x) || !(x > threshold));
bool predictedRejected = false;
DataTable table;
var results = ResultCollectionParameter.ActualValue;
if (!results.ContainsKey("RejectionCounts")) {
table = new DataTable("RejectionCounts");
results.Add(new Result("RejectionCounts", table));
var row = new DataRow("Predicted Rejected") { VisualProperties = { ChartType = DataRowVisualProperties.DataRowChartType.Histogram } };
table.Rows.Add(row);
row = new DataRow("Actually Rejected") { VisualProperties = { ChartType = DataRowVisualProperties.DataRowChartType.Histogram } };
table.Rows.Add(row);
// row = new DataRow("Actually Not Rejected") { VisualProperties = { ChartType = DataRowVisualProperties.DataRowChartType.Columns, StartIndexZero = true } };
// row.Values.AddRange(qualityPerInterval.Select(x => 0.0));
// table.Rows.Add(row);
//
// row = new DataRow("Predicted Not Rejected") { VisualProperties = { ChartType = DataRowVisualProperties.DataRowChartType.Columns, StartIndexZero = true } };
// row.Values.AddRange(qualityPerInterval.Select(x => 0.0));
// table.Rows.Add(row);
} else {
table = (DataTable)results["RejectionCounts"].Value;
}
i = 0;
foreach (var q in qualityPerInterval) {
if (double.IsNaN(q) || !(q > threshold)) {
predictedRejected = true;
break;
}
++i;
}
var actuallyRejected = !(actualQuality > parentQuality);
if (predictedRejected) {
table.Rows["Predicted Rejected"].Values.Add(i);
if (actuallyRejected)
table.Rows["Actually Rejected"].Values.Add(i);
}
// else {
// table.Rows["Predicted Not Rejected"].Values[i]++;
// if (!actuallyRejected)
// table.Rows["Actually Not Rejected"].Values[i]++;
// }
if (predictedRejected) {
if (actuallyRejected) {
CorrectlyRejected++;
} else {
IncorrectlyRejected++;
}
} else {
if (actuallyRejected) {
IncorrectlyNotRejected++;
} else {
CorrectlyNotRejected++;
}
}
return r * r;
}
public override double Evaluate(IExecutionContext context, ISymbolicExpressionTree tree, IRegressionProblemData problemData, IEnumerable rows) {
SymbolicDataAnalysisTreeInterpreterParameter.ExecutionContext = context;
EstimationLimitsParameter.ExecutionContext = context;
ApplyLinearScalingParameter.ExecutionContext = context;
var interpreter = SymbolicDataAnalysisTreeInterpreterParameter.ActualValue;
var estimationLimits = EstimationLimitsParameter.ActualValue;
var applyLinearScaling = ApplyLinearScalingParameter.ActualValue.Value;
double r2 = Calculate(interpreter, tree, estimationLimits.Lower, estimationLimits.Upper, problemData, rows, applyLinearScaling);
SymbolicDataAnalysisTreeInterpreterParameter.ExecutionContext = null;
EstimationLimitsParameter.ExecutionContext = null;
ApplyLinearScalingParameter.ExecutionContext = null;
return r2;
}
}
}