Changeset 15833 for branches/M5Regression/HeuristicLab.Algorithms.DataAnalysis/3.4

Timestamp:

03/08/18 10:57:03 (6 years ago)

Author:

bwerth

Message:

#2847: added handling of empty and underdetermined data sets

Location:

branches/M5Regression/HeuristicLab.Algorithms.DataAnalysis/3.4

Files:

• GBM/GradientBoostingRegressionAlgorithm.cs (modified) (12 diffs)
• M5Regression/LeafModels/PreconstructedLinearModel.cs (modified) (2 diffs)
• M5Regression/M5Regression.cs (modified) (4 diffs)

branches/M5Regression/HeuristicLab.Algorithms.DataAnalysis/3.4/GBM/GradientBoostingRegressionAlgorithm.cs

@@ -45 +45 @@
   [Creatable(CreatableAttribute.Categories.DataAnalysisRegression, Priority = 350)]
   public class GradientBoostingRegressionAlgorithm : FixedDataAnalysisAlgorithm<IRegressionProblem> {
-
     #region ParameterNames
-
     private const string IterationsParameterName = "Iterations";
     private const string NuParameterName = "Nu";
@@ -58 +56 @@
     private const string StoreRunsParameterName = "StoreRuns";
     private const string RegressionAlgorithmSolutionResultParameterName = "RegressionAlgorithmResult";
-
     #endregion
 
     #region ParameterProperties
-
     public IFixedValueParameter<IntValue> IterationsParameter {
       get { return (IFixedValueParameter<IntValue>)Parameters[IterationsParameterName]; }
@@ -102 +98 @@
       get { return (IFixedValueParameter<BoolValue>)Parameters[StoreRunsParameterName]; }
     }
-
     #endregion
 
     #region Properties
-
     public int Iterations {
       get { return IterationsParameter.Value.Value; }
@@ -155 +149 @@
       set { RegressionAlgorithmSolutionResultParameter.Value.Value = value; }
     }
-
     #endregion
 
     [StorableConstructor]
     protected GradientBoostingRegressionAlgorithm(bool deserializing)
-      : base(deserializing) {
-    }
+      : base(deserializing) { }
 
     protected GradientBoostingRegressionAlgorithm(GradientBoostingRegressionAlgorithm original, Cloner cloner)
-      : base(original, cloner) {
-    }
+      : base(original, cloner) { }
 
     public override IDeepCloneable Clone(Cloner cloner) {
@@ -232 +223 @@
       var problemData = Problem.ProblemData;
       var targetVarName = problemData.TargetVariable;
-      var activeVariables = problemData.AllowedInputVariables.Concat(new string[] { problemData.TargetVariable });
+      var activeVariables = problemData.AllowedInputVariables.Concat(new string[] {problemData.TargetVariable});
       var modifiableDataset = new ModifiableDataset(
         activeVariables,
@@ -252 +243 @@
       List<IRegressionModel> models = new List<IRegressionModel>();
       try {
-
         // Loop until iteration limit reached or canceled.
         for (int i = 0; i < Iterations; i++) {
@@ -258 +248 @@
 
           modifiableDataset.RemoveVariable(targetVarName);
-          modifiableDataset.AddVariable(targetVarName, curY.Concat(curYTest));
+          modifiableDataset.AddVariable(targetVarName, curY.Concat(curYTest).ToList());
 
           SampleTrainingData(rand, modifiableDataset, rRows, problemData.Dataset, curY, problemData.TargetVariable, problemData.TrainingIndices); // all training indices from the original problem data are allowed 
@@ -301 +291 @@
 
             models.Add(model);
-
-
           }
 
@@ -363 +351 @@
       alglib.lrunpack(lm, out coefficients, out features);
 
-      var ensembleModel = new RegressionEnsembleModel(models, coefficients.Take(models.Count)) { AverageModelEstimates = false };
+      var ensembleModel = new RegressionEnsembleModel(models, coefficients.Take(models.Count)) {AverageModelEstimates = false};
       var ensembleSolution = (IRegressionEnsembleSolution)ensembleModel.CreateRegressionSolution(problemData);
       return ensembleSolution;
@@ -442 +430 @@
         prob.ProblemDataParameter.Value = problemData;
         return true;
-      } else return false;
+      }
+      else return false;
     }
 
@@ -478 +467 @@
           // NaN evaluations would not be critical but are problematic if we want to combine all symbolic models into a single symbolic model 
           if (symbRegSol == null ||
-            (symbRegSol.TrainingLowerEstimationLimitHits == 0 && symbRegSol.TrainingUpperEstimationLimitHits == 0 &&
-             symbRegSol.TestLowerEstimationLimitHits == 0 && symbRegSol.TestUpperEstimationLimitHits == 0) &&
-            symbRegSol.TrainingNaNEvaluations == 0 && symbRegSol.TestNaNEvaluations == 0) {
+              (symbRegSol.TrainingLowerEstimationLimitHits == 0 && symbRegSol.TrainingUpperEstimationLimitHits == 0 &&
+               symbRegSol.TestLowerEstimationLimitHits == 0 && symbRegSol.TestUpperEstimationLimitHits == 0) &&
+              symbRegSol.TrainingNaNEvaluations == 0 && symbRegSol.TestNaNEvaluations == 0) {
             model = sol.Model;
           }
@@ -499 +488 @@
         ((BoolValue)paramItem.Parameters["SetSeedRandomly"].ActualValue).Value = false;
         ((IntValue)paramItem.Parameters["Seed"].ActualValue).Value = seed;
-      } else {
+      }
+      else {
         throw new ArgumentException("Base learner does not have a seed parameter (algorithm {0})", alg.Name);
       }
-
     }
   }

branches/M5Regression/HeuristicLab.Algorithms.DataAnalysis/3.4/M5Regression/LeafModels/PreconstructedLinearModel.cs

@@ -59 +59 @@
       SampleSize = original.SampleSize;
     }
-    public PreconstructedLinearModel(Dictionary<string, double> means, Dictionary<string, double> variances, Dictionary<string, double> coefficients, double intercept, string targetvariable) : base(targetvariable) {
+    public PreconstructedLinearModel(Dictionary<string, double> means, Dictionary<string, double> variances, Dictionary<string, double> coefficients, double intercept, string targetvariable, double residualVariance = 0, double sampleSize = 0) : base(targetvariable) {
       Coefficients = coefficients;
       Intercept = intercept;
@@ -178 +178 @@
       if (SampleSize == 0) return 0.0;
       var sum = (from var in Variances let d = dataset.GetDoubleValue(var.Key, row) - Means[var.Key] select d * d / var.Value).Sum();
-      var res = ResidualVariance * (1.0 / SampleSize + sum / (SampleSize - 1));
+      var res = ResidualVariance * (SampleSize - 1) / (SampleSize - 2) * (1.0 / SampleSize + sum / (SampleSize - 1));
       if (double.IsInfinity(res) || double.IsNaN(res)) return 0.0;
-      return res;
+      return Math.Sqrt(res);
     }
     #endregion

branches/M5Regression/HeuristicLab.Algorithms.DataAnalysis/3.4/M5Regression/M5Regression.cs

@@ -148 +148 @@
     #region Static Interface
     public static IRegressionSolution CreateRegressionSolution(IRegressionProblemData problemData, IRandom random, ILeafModel leafModel = null, ISplitter splitter = null, IPruning pruning = null,
-      bool useHoldout = false, double holdoutSize = 0.2, int minimumLeafSize = 4, bool generateRules = false, ResultCollection results = null, CancellationToken? cancellationToken = null) {
+      bool useHoldout = false, double holdoutSize = 0.2, int minimumLeafSize = 1, bool generateRules = false, ResultCollection results = null, CancellationToken? cancellationToken = null) {
       if (leafModel == null) leafModel = new LinearLeaf();
       if (splitter == null) splitter = new M5Splitter();
@@ -170 +170 @@
 
     #region Helpers
-    private static IScope InitializeScope(IRandom random, IRegressionProblemData problemData, IPruning pruning, int minLeafSize, ILeafModel leafModel, ISplitter splitter, bool GenerateRules, bool useHoldout, double holdoutSize) {
+    private static IScope InitializeScope(IRandom random, IRegressionProblemData problemData, IPruning pruning, int minLeafSize, ILeafModel leafModel, ISplitter splitter, bool generateRules, bool useHoldout, double holdoutSize) {
       var stateScope = new Scope("RegressionTreeStateScope");
 
@@ -196 +196 @@
       //store unbuilt model
       IItem model;
-      if (GenerateRules) {
+      if (generateRules) {
         model = RegressionRuleSetModel.CreateRuleModel(problemData.TargetVariable, regressionTreeParams);
         RegressionRuleSetModel.Initialize(stateScope);
@@ -215 +215 @@
 
     private static IRegressionModel Build(IScope stateScope, ResultCollection results, CancellationToken cancellationToken) {
+      var regressionTreeParams = (RegressionTreeParameters)stateScope.Variables[RegressionTreeParameterVariableName].Value;
       var model = (IM5Model)stateScope.Variables[ModelVariableName].Value;
       var trainingRows = (IntArray)stateScope.Variables[TrainingSetVariableName].Value;
       var pruningRows = (IntArray)stateScope.Variables[PruningSetVariableName].Value;
+      if (1 > trainingRows.Length)
+        return new PreconstructedLinearModel(new Dictionary<string, double>(), new Dictionary<string, double>(), new Dictionary<string, double>(), 0, regressionTreeParams.TargetVariable);
+      if (regressionTreeParams.MinLeafSize > trainingRows.Length) {
+        var targets = regressionTreeParams.Data.GetDoubleValues(regressionTreeParams.TargetVariable).ToArray();
+        return new PreconstructedLinearModel(new Dictionary<string, double>(), new Dictionary<string, double>(), new Dictionary<string, double>(), targets.Average(), regressionTreeParams.TargetVariable, targets.Variance(), targets.Length);
+      }
       model.Build(trainingRows.ToArray(), pruningRows.ToArray(), stateScope, results, cancellationToken);
       return model;