Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2994-AutoDiffForIntervals/HeuristicLab.Problems.DataAnalysis.Regression.Symbolic.Extensions/ConstrainedConstantOptimizationEvaluator.cs @ 16941

Last change on this file since 16941 was 16941, checked in by gkronber, 5 years ago

#2994: make branch compile with NamedIntervals branch, disable alglib optguard, copylocal = false

File size: 27.3 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2019 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
22using System;
23using System.Collections.Generic;
24using System.Linq;
25using HeuristicLab.Common;
26using HeuristicLab.Core;
27using HeuristicLab.Data;
28using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
29using HeuristicLab.Optimization;
30using HeuristicLab.Parameters;
31using HEAL.Attic;
32
33namespace HeuristicLab.Problems.DataAnalysis.Symbolic.Regression {
34  [Item("Constant Optimization Evaluator (with constraints)", "")]
35  [StorableType("A8958E06-C54A-4193-862E-8315C86EB5C1")]
36  public class ConstrainedConstantOptimizationEvaluator : SymbolicRegressionSingleObjectiveEvaluator {
37    private const string ConstantOptimizationIterationsParameterName = "ConstantOptimizationIterations";
38    private const string ConstantOptimizationImprovementParameterName = "ConstantOptimizationImprovement";
39    private const string ConstantOptimizationProbabilityParameterName = "ConstantOptimizationProbability";
40    private const string ConstantOptimizationRowsPercentageParameterName = "ConstantOptimizationRowsPercentage";
41    private const string UpdateConstantsInTreeParameterName = "UpdateConstantsInSymbolicExpressionTree";
42    private const string UpdateVariableWeightsParameterName = "Update Variable Weights";
43
44    private const string FunctionEvaluationsResultParameterName = "Constants Optimization Function Evaluations";
45    private const string GradientEvaluationsResultParameterName = "Constants Optimization Gradient Evaluations";
46    private const string CountEvaluationsParameterName = "Count Function and Gradient Evaluations";
47
48    public IFixedValueParameter<IntValue> ConstantOptimizationIterationsParameter {
49      get { return (IFixedValueParameter<IntValue>)Parameters[ConstantOptimizationIterationsParameterName]; }
50    }
51    public IFixedValueParameter<DoubleValue> ConstantOptimizationImprovementParameter {
52      get { return (IFixedValueParameter<DoubleValue>)Parameters[ConstantOptimizationImprovementParameterName]; }
53    }
54    public IFixedValueParameter<PercentValue> ConstantOptimizationProbabilityParameter {
55      get { return (IFixedValueParameter<PercentValue>)Parameters[ConstantOptimizationProbabilityParameterName]; }
56    }
57    public IFixedValueParameter<PercentValue> ConstantOptimizationRowsPercentageParameter {
58      get { return (IFixedValueParameter<PercentValue>)Parameters[ConstantOptimizationRowsPercentageParameterName]; }
59    }
60    public IFixedValueParameter<BoolValue> UpdateConstantsInTreeParameter {
61      get { return (IFixedValueParameter<BoolValue>)Parameters[UpdateConstantsInTreeParameterName]; }
62    }
63    public IFixedValueParameter<BoolValue> UpdateVariableWeightsParameter {
64      get { return (IFixedValueParameter<BoolValue>)Parameters[UpdateVariableWeightsParameterName]; }
65    }
66
67    public IResultParameter<IntValue> FunctionEvaluationsResultParameter {
68      get { return (IResultParameter<IntValue>)Parameters[FunctionEvaluationsResultParameterName]; }
69    }
70    public IResultParameter<IntValue> GradientEvaluationsResultParameter {
71      get { return (IResultParameter<IntValue>)Parameters[GradientEvaluationsResultParameterName]; }
72    }
73    public IFixedValueParameter<BoolValue> CountEvaluationsParameter {
74      get { return (IFixedValueParameter<BoolValue>)Parameters[CountEvaluationsParameterName]; }
75    }
76
77
78    public IntValue ConstantOptimizationIterations {
79      get { return ConstantOptimizationIterationsParameter.Value; }
80    }
81    public DoubleValue ConstantOptimizationImprovement {
82      get { return ConstantOptimizationImprovementParameter.Value; }
83    }
84    public PercentValue ConstantOptimizationProbability {
85      get { return ConstantOptimizationProbabilityParameter.Value; }
86    }
87    public PercentValue ConstantOptimizationRowsPercentage {
88      get { return ConstantOptimizationRowsPercentageParameter.Value; }
89    }
90    public bool UpdateConstantsInTree {
91      get { return UpdateConstantsInTreeParameter.Value.Value; }
92      set { UpdateConstantsInTreeParameter.Value.Value = value; }
93    }
94
95    public bool UpdateVariableWeights {
96      get { return UpdateVariableWeightsParameter.Value.Value; }
97      set { UpdateVariableWeightsParameter.Value.Value = value; }
98    }
99
100    public bool CountEvaluations {
101      get { return CountEvaluationsParameter.Value.Value; }
102      set { CountEvaluationsParameter.Value.Value = value; }
103    }
104
105    public override bool Maximization {
106      get { return false; }
107    }
108
109    [StorableConstructor]
110    protected ConstrainedConstantOptimizationEvaluator(StorableConstructorFlag _) : base(_) { }
111    protected ConstrainedConstantOptimizationEvaluator(ConstrainedConstantOptimizationEvaluator original, Cloner cloner)
112      : base(original, cloner) {
113    }
114    public ConstrainedConstantOptimizationEvaluator()
115      : base() {
116      Parameters.Add(new FixedValueParameter<IntValue>(ConstantOptimizationIterationsParameterName, "Determines how many iterations should be calculated while optimizing the constant of a symbolic expression tree (0 indicates other or default stopping criterion).", new IntValue(10)));
117      Parameters.Add(new FixedValueParameter<DoubleValue>(ConstantOptimizationImprovementParameterName, "Determines the relative improvement which must be achieved in the constant optimization to continue with it (0 indicates other or default stopping criterion).", new DoubleValue(0)) { Hidden = true });
118      Parameters.Add(new FixedValueParameter<PercentValue>(ConstantOptimizationProbabilityParameterName, "Determines the probability that the constants are optimized", new PercentValue(1)));
119      Parameters.Add(new FixedValueParameter<PercentValue>(ConstantOptimizationRowsPercentageParameterName, "Determines the percentage of the rows which should be used for constant optimization", new PercentValue(1)));
120      Parameters.Add(new FixedValueParameter<BoolValue>(UpdateConstantsInTreeParameterName, "Determines if the constants in the tree should be overwritten by the optimized constants.", new BoolValue(true)) { Hidden = true });
121      Parameters.Add(new FixedValueParameter<BoolValue>(UpdateVariableWeightsParameterName, "Determines if the variable weights in the tree should be  optimized.", new BoolValue(true)) { Hidden = true });
122
123      Parameters.Add(new FixedValueParameter<BoolValue>(CountEvaluationsParameterName, "Determines if function and gradient evaluation should be counted.", new BoolValue(false)));
124      Parameters.Add(new ResultParameter<IntValue>(FunctionEvaluationsResultParameterName, "The number of function evaluations performed by the constants optimization evaluator", "Results", new IntValue()));
125      Parameters.Add(new ResultParameter<IntValue>(GradientEvaluationsResultParameterName, "The number of gradient evaluations performed by the constants optimization evaluator", "Results", new IntValue()));
126    }
127
128    public override IDeepCloneable Clone(Cloner cloner) {
129      return new ConstrainedConstantOptimizationEvaluator(this, cloner);
130    }
131
132    [StorableHook(HookType.AfterDeserialization)]
133    private void AfterDeserialization() { }
134
135    private static readonly object locker = new object();
136
137    public override IOperation InstrumentedApply() {
138      var solution = SymbolicExpressionTreeParameter.ActualValue;
139      double quality;
140      if (RandomParameter.ActualValue.NextDouble() < ConstantOptimizationProbability.Value) {
141        IEnumerable<int> constantOptimizationRows = GenerateRowsToEvaluate(ConstantOptimizationRowsPercentage.Value);
142        var counter = new EvaluationsCounter();
143        quality = OptimizeConstants(SymbolicDataAnalysisTreeInterpreterParameter.ActualValue, solution, ProblemDataParameter.ActualValue,
144           constantOptimizationRows, ApplyLinearScalingParameter.ActualValue.Value, ConstantOptimizationIterations.Value, updateVariableWeights: UpdateVariableWeights, lowerEstimationLimit: EstimationLimitsParameter.ActualValue.Lower, upperEstimationLimit: EstimationLimitsParameter.ActualValue.Upper, updateConstantsInTree: UpdateConstantsInTree, counter: counter);
145
146        if (ConstantOptimizationRowsPercentage.Value != RelativeNumberOfEvaluatedSamplesParameter.ActualValue.Value) {
147          var evaluationRows = GenerateRowsToEvaluate();
148          quality = SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(SymbolicDataAnalysisTreeInterpreterParameter.ActualValue, solution, double.MinValue, double.MaxValue, ProblemDataParameter.ActualValue, evaluationRows, applyLinearScaling: false);
149        }
150
151        if (CountEvaluations) {
152          lock (locker) {
153            FunctionEvaluationsResultParameter.ActualValue.Value += counter.FunctionEvaluations;
154            GradientEvaluationsResultParameter.ActualValue.Value += counter.GradientEvaluations;
155          }
156        }
157
158      } else {
159        var evaluationRows = GenerateRowsToEvaluate();
160        quality = SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(SymbolicDataAnalysisTreeInterpreterParameter.ActualValue, solution, double.MinValue, double.MaxValue, ProblemDataParameter.ActualValue, evaluationRows, applyLinearScaling: false);
161      }
162      QualityParameter.ActualValue = new DoubleValue(quality);
163
164      return base.InstrumentedApply();
165    }
166
167    public override double Evaluate(IExecutionContext context, ISymbolicExpressionTree tree, IRegressionProblemData problemData, IEnumerable<int> rows) {
168      SymbolicDataAnalysisTreeInterpreterParameter.ExecutionContext = context;
169      EstimationLimitsParameter.ExecutionContext = context;
170      ApplyLinearScalingParameter.ExecutionContext = context;
171      FunctionEvaluationsResultParameter.ExecutionContext = context;
172      GradientEvaluationsResultParameter.ExecutionContext = context;
173
174      // MSE evaluator is used on purpose instead of the const-opt evaluator,
175      // because Evaluate() is used to get the quality of evolved models on
176      // different partitions of the dataset (e.g., best validation model)
177      double mse = SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(SymbolicDataAnalysisTreeInterpreterParameter.ActualValue, tree, double.MinValue, double.MaxValue, problemData, rows, applyLinearScaling: false);
178
179      SymbolicDataAnalysisTreeInterpreterParameter.ExecutionContext = null;
180      EstimationLimitsParameter.ExecutionContext = null;
181      ApplyLinearScalingParameter.ExecutionContext = null;
182      FunctionEvaluationsResultParameter.ExecutionContext = null;
183      GradientEvaluationsResultParameter.ExecutionContext = null;
184
185      return mse;
186    }
187
188    public class EvaluationsCounter {
189      public int FunctionEvaluations = 0;
190      public int GradientEvaluations = 0;
191    }
192
193    private static void GetParameterNodes(ISymbolicExpressionTree tree, out List<ISymbolicExpressionTreeNode> thetaNodes, out List<double> thetaValues) {
194      thetaNodes = new List<ISymbolicExpressionTreeNode>();
195      thetaValues = new List<double>();
196
197      var nodes = tree.IterateNodesPrefix().ToArray();
198      for (int i = 0; i < nodes.Length; ++i) {
199        var node = nodes[i];
200        if (node is VariableTreeNode variableTreeNode) {
201          thetaValues.Add(variableTreeNode.Weight);
202          thetaNodes.Add(node);
203        } else if (node is ConstantTreeNode constantTreeNode) {
204          thetaNodes.Add(node);
205          thetaValues.Add(constantTreeNode.Value);
206        }
207      }
208    }
209
210    public static double OptimizeConstants(ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
211      ISymbolicExpressionTree tree, IRegressionProblemData problemData, IEnumerable<int> rows, bool applyLinearScaling,
212      int maxIterations, bool updateVariableWeights = true,
213      double lowerEstimationLimit = double.MinValue, double upperEstimationLimit = double.MaxValue,
214      bool updateConstantsInTree = true, Action<double[], double, object> iterationCallback = null, EvaluationsCounter counter = null) {
215
216      if (!updateVariableWeights) throw new NotSupportedException("not updating variable weights is not supported");
217      if (!updateConstantsInTree) throw new NotSupportedException("not updating tree parameters is not supported");
218      if (applyLinearScaling) throw new NotSupportedException("linear scaling is not supported");
219
220      // we always update constants, so we don't need to calculate initial quality
221      // double originalQuality = SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(interpreter, tree, lowerEstimationLimit, upperEstimationLimit, problemData, rows, applyLinearScaling: false);
222
223      if (counter == null) counter = new EvaluationsCounter();
224      var rowEvaluationsCounter = new EvaluationsCounter();
225
226      var intervalConstraints = problemData.IntervalConstraints;
227      var dataIntervals = problemData.VariableRanges.GetIntervals();
228
229      // convert constants to variables named theta...
230      var treeForDerivation = ReplaceConstWithVar(tree, out List<string> thetaNames, out List<double> thetaValues); // copies the tree
231
232      // create trees for relevant derivatives
233      Dictionary<string, ISymbolicExpressionTree> derivatives = new Dictionary<string, ISymbolicExpressionTree>();
234      var allThetaNodes = thetaNames.Select(_ => new List<ConstantTreeNode>()).ToArray();
235      var constraintTrees = new List<ISymbolicExpressionTree>();
236      foreach (var constraint in intervalConstraints.Constraints) {
237        if (constraint.IsDerivation) {
238          if (!problemData.AllowedInputVariables.Contains(constraint.Variable))
239            throw new ArgumentException($"Invalid constraint: the variable {constraint.Variable} does not exist in the dataset.");
240          var df = DerivativeCalculator.Derive(treeForDerivation, constraint.Variable);
241
242          // alglib requires constraint expressions of the form c(x) <= 0
243          // -> we make two expressions, one for the lower bound and one for the upper bound
244
245          if (constraint.Interval.UpperBound < double.PositiveInfinity) {
246            var df_smaller_upper = Subtract((ISymbolicExpressionTree)df.Clone(), CreateConstant(constraint.Interval.UpperBound));
247            // convert variables named theta back to constants
248            var df_prepared = ReplaceVarWithConst(df_smaller_upper, thetaNames, thetaValues, allThetaNodes);
249            constraintTrees.Add(df_prepared);
250          }
251          if (constraint.Interval.LowerBound > double.NegativeInfinity) {
252            var df_larger_lower = Subtract(CreateConstant(constraint.Interval.LowerBound), (ISymbolicExpressionTree)df.Clone());
253            // convert variables named theta back to constants
254            var df_prepared = ReplaceVarWithConst(df_larger_lower, thetaNames, thetaValues, allThetaNodes);
255            constraintTrees.Add(df_prepared);
256          }
257        } else {
258          if (constraint.Interval.UpperBound < double.PositiveInfinity) {
259            var f_smaller_upper = Subtract((ISymbolicExpressionTree)treeForDerivation.Clone(), CreateConstant(constraint.Interval.UpperBound));
260            // convert variables named theta back to constants
261            var df_prepared = ReplaceVarWithConst(f_smaller_upper, thetaNames, thetaValues, allThetaNodes);
262            constraintTrees.Add(df_prepared);
263          }
264          if (constraint.Interval.LowerBound > double.NegativeInfinity) {
265            var f_larger_lower = Subtract(CreateConstant(constraint.Interval.LowerBound), (ISymbolicExpressionTree)treeForDerivation.Clone());
266            // convert variables named theta back to constants
267            var df_prepared = ReplaceVarWithConst(f_larger_lower, thetaNames, thetaValues, allThetaNodes);
268            constraintTrees.Add(df_prepared);
269          }
270        }
271      }
272
273      var preparedTree = ReplaceVarWithConst(treeForDerivation, thetaNames, thetaValues, allThetaNodes);
274
275
276      // local function
277      void UpdateThetaValues(double[] theta) {
278        for (int i = 0; i < theta.Length; ++i) {
279          foreach (var constNode in allThetaNodes[i]) constNode.Value = theta[i];
280        }
281      }
282
283      // buffers for calculate_jacobian
284      var target = problemData.TargetVariableTrainingValues.ToArray();
285      var fi_eval = new double[target.Length];
286      var jac_eval = new double[target.Length, thetaValues.Count];
287
288      // define the callback used by the alglib optimizer
289      // the x argument for this callback represents our theta
290      // local function
291      void calculate_jacobian(double[] x, double[] fi, double[,] jac, object obj) {
292        UpdateThetaValues(x);
293
294        var autoDiffEval = new VectorAutoDiffEvaluator();
295        autoDiffEval.Evaluate(preparedTree, problemData.Dataset, problemData.TrainingIndices.ToArray(),
296          GetParameterNodes(preparedTree, allThetaNodes), fi_eval, jac_eval);
297
298        // calc sum of squared errors and gradient
299        var sse = 0.0;
300        var g = new double[x.Length];
301        for (int i = 0; i < target.Length; i++) {
302          var res = target[i] - fi_eval[i];
303          sse += 0.5 * res * res;
304          for (int j = 0; j < g.Length; j++) {
305            g[j] -= res * jac_eval[i, j];
306          }
307        }
308
309        fi[0] = sse / target.Length;
310        for (int j = 0; j < x.Length; j++) { jac[0, j] = g[j] / target.Length; }
311
312        var intervalEvaluator = new IntervalEvaluator();
313        for (int i = 0; i < constraintTrees.Count; i++) {
314          var interval = intervalEvaluator.Evaluate(constraintTrees[i], dataIntervals, GetParameterNodes(constraintTrees[i], allThetaNodes),
315            out double[] lowerGradient, out double[] upperGradient);
316
317          // we transformed this to a constraint c(x) <= 0, so only the upper bound is relevant for us
318          fi[i + 1] = interval.UpperBound;
319          for (int j = 0; j < x.Length; j++) {
320            jac[i + 1, j] = upperGradient[j];
321          }
322        }
323      }
324
325
326
327      alglib.minnlcstate state;
328      alglib.minnlcreport rep;
329      try {
330        alglib.minnlccreate(thetaValues.Count, thetaValues.ToArray(), out state);
331        alglib.minnlcsetalgoslp(state);        // SLP is more robust but slower
332        alglib.minnlcsetbc(state, thetaValues.Select(_ => -10000.0).ToArray(), thetaValues.Select(_ => +10000.0).ToArray());
333        alglib.minnlcsetcond(state, 1E-7, maxIterations);
334        var s = Enumerable.Repeat(1d, thetaValues.Count).ToArray();  // scale is set to unit scale
335        alglib.minnlcsetscale(state, s);
336
337        // set non-linear constraints: 0 equality constraints, constraintTrees inequality constraints
338        alglib.minnlcsetnlc(state, 0, constraintTrees.Count);
339
340        alglib.minnlcoptimize(state, calculate_jacobian, null, null);
341        alglib.minnlcresults(state, out double[] xOpt, out rep);
342
343
344        // counter.FunctionEvaluations += rep.nfev; TODO
345        counter.GradientEvaluations += rep.nfev;
346
347        if (rep.terminationtype != -8) {
348          // update parameters in tree
349          var pIdx = 0;
350          foreach (var node in tree.IterateNodesPostfix()) {
351            if(node is ConstantTreeNode constTreeNode) {
352              constTreeNode.Value = xOpt[pIdx++];
353            } else if(node is VariableTreeNode varTreeNode) {
354              varTreeNode.Weight = xOpt[pIdx++];
355            }
356          }
357
358          // note: we keep the optimized constants even when the tree is worse.
359        }
360
361      } catch (ArithmeticException) {
362        // eval MSE of original tree
363        return SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(interpreter, tree, lowerEstimationLimit, upperEstimationLimit, problemData, rows, applyLinearScaling: false);
364
365      } catch (alglib.alglibexception) {
366        // eval MSE of original tree
367        return SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(interpreter, tree, lowerEstimationLimit, upperEstimationLimit, problemData, rows, applyLinearScaling: false);
368      }
369
370
371      // evaluate tree with updated constants
372      return SymbolicRegressionSingleObjectiveMeanSquaredErrorEvaluator.Calculate(interpreter, tree, lowerEstimationLimit, upperEstimationLimit, problemData, rows, applyLinearScaling: false);
373    }
374
375    #region helper
376    private static ISymbolicExpressionTreeNode[] GetParameterNodes(ISymbolicExpressionTree tree, List<ConstantTreeNode>[] allNodes) {
377      // TODO better solution necessary
378      var treeConstNodes = tree.IterateNodesPostfix().OfType<ConstantTreeNode>().ToArray();
379      var paramNodes = new ISymbolicExpressionTreeNode[allNodes.Length];
380      for (int i = 0; i < paramNodes.Length; i++) {
381        paramNodes[i] = allNodes[i].SingleOrDefault(n => treeConstNodes.Contains(n));
382      }
383      return paramNodes;
384    }
385
386    private static ISymbolicExpressionTree ReplaceVarWithConst(ISymbolicExpressionTree tree, List<string> thetaNames, List<double> thetaValues, List<ConstantTreeNode>[] thetaNodes) {
387      var copy = (ISymbolicExpressionTree)tree.Clone();
388      var nodes = copy.IterateNodesPostfix().ToList();
389      for (int i = 0; i < nodes.Count; i++) {
390        var n = nodes[i] as VariableTreeNode;
391        if (n != null) {
392          var thetaIdx = thetaNames.IndexOf(n.VariableName);
393          if (thetaIdx >= 0) {
394            var parent = n.Parent;
395            if (thetaNodes[thetaIdx].Any()) {
396              // HACK: REUSE CONSTANT TREE NODE IN SEVERAL TREES
397              // we use this trick to allow autodiff over thetas when thetas occurr multiple times in the tree (e.g. in derived trees)
398              var constNode = thetaNodes[thetaIdx].First();
399              var childIdx = parent.IndexOfSubtree(n);
400              parent.RemoveSubtree(childIdx);
401              parent.InsertSubtree(childIdx, constNode);
402            } else {
403              var constNode = (ConstantTreeNode)CreateConstant(thetaValues[thetaIdx]);
404              var childIdx = parent.IndexOfSubtree(n);
405              parent.RemoveSubtree(childIdx);
406              parent.InsertSubtree(childIdx, constNode);
407              thetaNodes[thetaIdx].Add(constNode);
408            }
409          }
410        }
411      }
412      return copy;
413    }
414
415    private static ISymbolicExpressionTree ReplaceConstWithVar(ISymbolicExpressionTree tree, out List<string> thetaNames, out List<double> thetaValues) {
416      thetaNames = new List<string>();
417      thetaValues = new List<double>();
418      var copy = (ISymbolicExpressionTree)tree.Clone();
419      var nodes = copy.IterateNodesPostfix().ToList();
420
421      int n = 1;
422      for (int i = 0; i < nodes.Count; ++i) {
423        var node = nodes[i];
424        if (node is ConstantTreeNode constantTreeNode) {
425          var thetaVar = (VariableTreeNode)new Problems.DataAnalysis.Symbolic.Variable().CreateTreeNode();
426          thetaVar.Weight = 1;
427          thetaVar.VariableName = $"θ{n++}";
428
429          thetaNames.Add(thetaVar.VariableName);
430          thetaValues.Add(constantTreeNode.Value);
431
432          var parent = constantTreeNode.Parent;
433          if (parent != null) {
434            var index = constantTreeNode.Parent.IndexOfSubtree(constantTreeNode);
435            parent.RemoveSubtree(index);
436            parent.InsertSubtree(index, thetaVar);
437          }
438        }
439        if (node is VariableTreeNode varTreeNode) {
440          var thetaVar = (VariableTreeNode)new Problems.DataAnalysis.Symbolic.Variable().CreateTreeNode();
441          thetaVar.Weight = 1;
442          thetaVar.VariableName = $"θ{n++}";
443
444          thetaNames.Add(thetaVar.VariableName);
445          thetaValues.Add(varTreeNode.Weight);
446
447          var parent = varTreeNode.Parent;
448          if (parent != null) {
449            var index = varTreeNode.Parent.IndexOfSubtree(varTreeNode);
450            parent.RemoveSubtree(index);
451            var prodNode = MakeNode<Multiplication>();
452            varTreeNode.Weight = 1.0;
453            prodNode.AddSubtree(varTreeNode);
454            prodNode.AddSubtree(thetaVar);
455            parent.InsertSubtree(index, prodNode);
456          }
457        }
458      }
459      return copy;
460    }
461
462    private static ISymbolicExpressionTreeNode CreateConstant(double value) {
463      var constantNode = (ConstantTreeNode)new Constant().CreateTreeNode();
464      constantNode.Value = value;
465      return constantNode;
466    }
467
468    private static ISymbolicExpressionTree Subtract(ISymbolicExpressionTree t, ISymbolicExpressionTreeNode b) {
469      var sub = MakeNode<Subtraction>(t.Root.GetSubtree(0).GetSubtree(0), b);
470      t.Root.GetSubtree(0).RemoveSubtree(0);
471      t.Root.GetSubtree(0).InsertSubtree(0, sub);
472      return t;
473    }
474    private static ISymbolicExpressionTree Subtract(ISymbolicExpressionTreeNode b, ISymbolicExpressionTree t) {
475      var sub = MakeNode<Subtraction>(b, t.Root.GetSubtree(0).GetSubtree(0));
476      t.Root.GetSubtree(0).RemoveSubtree(0);
477      t.Root.GetSubtree(0).InsertSubtree(0, sub);
478      return t;
479    }
480
481    private static ISymbolicExpressionTreeNode MakeNode<T>(params ISymbolicExpressionTreeNode[] fs) where T : ISymbol, new() {
482      var node = new T().CreateTreeNode();
483      foreach (var f in fs) node.AddSubtree(f);
484      return node;
485    }
486    #endregion
487
488    private static void UpdateConstants(ISymbolicExpressionTreeNode[] nodes, double[] constants) {
489      if (nodes.Length != constants.Length) throw new InvalidOperationException();
490      for (int i = 0; i < nodes.Length; i++) {
491        if (nodes[i] is VariableTreeNode varNode) varNode.Weight = constants[i];
492        else if (nodes[i] is ConstantTreeNode constNode) constNode.Value = constants[i];
493      }
494    }
495
496    private static alglib.ndimensional_fvec CreateFunc(ISymbolicExpressionTree tree, VectorEvaluator eval, ISymbolicExpressionTreeNode[] parameterNodes, IDataset ds, string targetVar, int[] rows) {
497      var y = ds.GetDoubleValues(targetVar, rows).ToArray();
498      return (double[] c, double[] fi, object o) => {
499        UpdateConstants(parameterNodes, c);
500        var pred = eval.Evaluate(tree, ds, rows);
501        for (int i = 0; i < fi.Length; i++)
502          fi[i] = pred[i] - y[i];
503
504        var counter = (EvaluationsCounter)o;
505        counter.FunctionEvaluations++;
506      };
507    }
508
509    private static alglib.ndimensional_jac CreateJac(ISymbolicExpressionTree tree, VectorAutoDiffEvaluator eval, ISymbolicExpressionTreeNode[] parameterNodes, IDataset ds, string targetVar, int[] rows) {
510      var y = ds.GetDoubleValues(targetVar, rows).ToArray();
511      return (double[] c, double[] fi, double[,] jac, object o) => {
512        UpdateConstants(parameterNodes, c);
513        eval.Evaluate(tree, ds, rows, parameterNodes, fi, jac);
514
515        for (int i = 0; i < fi.Length; i++)
516          fi[i] -= y[i];
517
518        var counter = (EvaluationsCounter)o;
519        counter.GradientEvaluations++;
520      };
521    }
522    public static bool CanOptimizeConstants(ISymbolicExpressionTree tree) {
523      return TreeToAutoDiffTermConverter.IsCompatible(tree);
524    }
525  }
526}
Note: See TracBrowser for help on using the repository browser.