Changeset 16141 for branches/2817-BinPackingSpeedup/HeuristicLab.Problems.DataAnalysis/3.4/Implementation

Timestamp:

09/14/18 11:47:37 (6 years ago)

Author:

abeham

Message:

#2817: updated to trunk r16140

Location:

branches/2817-BinPackingSpeedup

Files:

• . (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis (modified) (1 prop)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationSolutionVariableImpactsCalculator.cs (modified) (6 diffs)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/DataAnalysisProblemData.cs (modified) (1 diff)
• HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionSolutionVariableImpactsCalculator.cs (modified) (4 diffs)

branches/2817-BinPackingSpeedup/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Classification/ClassificationSolutionVariableImpactsCalculator.cs

@@ -100 +100 @@
       var problemData = solution.ProblemData;
       var dataset = problemData.Dataset;
+      var model = (IClassificationModel)solution.Model.Clone(); //mkommend: clone of model is necessary, because the thresholds for IDiscriminantClassificationModels are updated
 
       IEnumerable<int> rows;
@@ -137 +138 @@
       // calculate impacts for double variables
       foreach (var inputVariable in allowedInputVariables.Where(problemData.Dataset.VariableHasType<double>)) {
-        var newEstimates = EvaluateModelWithReplacedVariable(solution.Model, inputVariable, modifiableDataset, rows, replacementMethod);
+        var newEstimates = EvaluateModelWithReplacedVariable(model, inputVariable, modifiableDataset, rows, replacementMethod);
         var newAccuracy = OnlineAccuracyCalculator.Calculate(targetValues, newEstimates, out error);
         if (error != OnlineCalculatorError.None) throw new InvalidOperationException("Error during R² calculation with replaced inputs.");
@@ -150 +151 @@
           var smallestImpact = double.PositiveInfinity;
           foreach (var repl in problemData.Dataset.GetStringValues(inputVariable, rows).Distinct()) {
-            var newEstimates = EvaluateModelWithReplacedVariable(solution.Model, inputVariable, modifiableDataset, rows,
+            var newEstimates = EvaluateModelWithReplacedVariable(model, inputVariable, modifiableDataset, rows,
               Enumerable.Repeat(repl, dataset.Rows));
             var newAccuracy = OnlineAccuracyCalculator.Calculate(targetValues, newEstimates, out error);
@@ -164 +165 @@
           // calculate impacts for factor variables
 
-          var newEstimates = EvaluateModelWithReplacedVariable(solution.Model, inputVariable, modifiableDataset, rows,
+          var newEstimates = EvaluateModelWithReplacedVariable(model, inputVariable, modifiableDataset, rows,
             factorReplacementMethod);
           var newAccuracy = OnlineAccuracyCalculator.Calculate(targetValues, newEstimates, out error);
@@ -263 +264 @@
       var originalValues = dataset.GetReadOnlyDoubleValues(variable).ToList();
       dataset.ReplaceVariable(variable, replacementValues.ToList());
+
+      var discModel = model as IDiscriminantFunctionClassificationModel;
+      if (discModel != null) {
+        var problemData = new ClassificationProblemData(dataset, dataset.VariableNames, model.TargetVariable);
+        discModel.RecalculateModelParameters(problemData, rows);
+      }
+
       //mkommend: ToList is used on purpose to avoid lazy evaluation that could result in wrong estimates due to variable replacements
       var estimates = model.GetEstimatedClassValues(dataset, rows).ToList();
@@ -273 +281 @@
       var originalValues = dataset.GetReadOnlyStringValues(variable).ToList();
       dataset.ReplaceVariable(variable, replacementValues.ToList());
+
+
+      var discModel = model as IDiscriminantFunctionClassificationModel;
+      if (discModel != null) {
+        var problemData = new ClassificationProblemData(dataset, dataset.VariableNames, model.TargetVariable);
+        discModel.RecalculateModelParameters(problemData, rows);
+      }
+
       //mkommend: ToList is used on purpose to avoid lazy evaluation that could result in wrong estimates due to variable replacements
       var estimates = model.GetEstimatedClassValues(dataset, rows).ToList();

branches/2817-BinPackingSpeedup/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/DataAnalysisProblemData.cs

@@ -163 +163 @@
 
       var variables = dataset.VariableNames.Where(variable => dataset.VariableHasType<double>(variable) || dataset.VariableHasType<string>(variable));
-      var inputVariables = new CheckedItemList<StringValue>(variables.Select(x => new StringValue(x)));
+      var inputVariables = new CheckedItemList<StringValue>(variables.Select(x => new StringValue(x).AsReadOnly()));
       foreach (StringValue x in inputVariables)
         inputVariables.SetItemCheckedState(x, allowedInputVariables.Contains(x.Value));

branches/2817-BinPackingSpeedup/HeuristicLab.Problems.DataAnalysis/3.4/Implementation/Regression/RegressionSolutionVariableImpactsCalculator.cs

@@ -52 +52 @@
       All
     }
-   
+
     private const string ReplacementParameterName = "Replacement Method";
     private const string DataPartitionParameterName = "DataPartition";
@@ -96 +96 @@
       DataPartitionEnum data = DataPartitionEnum.Training,
       ReplacementMethodEnum replacementMethod = ReplacementMethodEnum.Median,
-      FactorReplacementMethodEnum factorReplacementMethod = FactorReplacementMethodEnum.Best) {
+      FactorReplacementMethodEnum factorReplacementMethod = FactorReplacementMethodEnum.Best,
+      Func<double, string, bool> progressCallback = null) {
 
       var problemData = solution.ProblemData;
@@ -134 +135 @@
       var allowedInputVariables = dataset.VariableNames.Where(v => inputvariables.Contains(v)).ToList();
 
+      int curIdx = 0;
+      int count = allowedInputVariables.Where(problemData.Dataset.VariableHasType<double>).Count();
       // calculate impacts for double variables
       foreach (var inputVariable in allowedInputVariables.Where(problemData.Dataset.VariableHasType<double>)) {
+        //Report the current progress in percent. If the callback returns true, it means the execution shall be stopped
+        if (progressCallback != null) {
+          curIdx++;
+          if (progressCallback((double)curIdx / count, string.Format("Calculating impact for variable {0} ({1} of {2})", inputVariable, curIdx, count))) { return null; }
+        }
         var newEstimates = EvaluateModelWithReplacedVariable(solution.Model, inputVariable, modifiableDataset, rows, replacementMethod);
         var newR2 = OnlinePearsonsRCalculator.Calculate(targetValues, newEstimates, out error);
@@ -180 +188 @@
     }
 
+
     private static IEnumerable<double> EvaluateModelWithReplacedVariable(IRegressionModel model, string variable, ModifiableDataset dataset, IEnumerable<int> rows, ReplacementMethodEnum replacement = ReplacementMethodEnum.Median) {
       var originalValues = dataset.GetReadOnlyDoubleValues(variable).ToList();