Changeset 10908 for branches/DataPreprocessing/HeuristicLab.DataPreprocessing/3.3/Implementations/ChartLogic.cs

Timestamp:

05/28/14 15:09:26 (10 years ago)

Author:

mleitner

Message:

Add Feature correlation matrix, Add limit for distinct values in histogramm classification.

File:

• branches/DataPreprocessing/HeuristicLab.DataPreprocessing/3.3/Implementations/ChartLogic.cs (modified) (5 diffs)

branches/DataPreprocessing/HeuristicLab.DataPreprocessing/3.3/Implementations/ChartLogic.cs

@@ -1 +1 @@
 #region License Information
 /* HeuristicLab
- * Copyright (C) 2002-2013 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
+ * Copyright (C) 2002-2014 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
  *
  * This file is part of HeuristicLab.
@@ -32 +32 @@
 
   public class ChartLogic : IChartLogic {
- 
+    private const int MAX_DISTINCT_VALUES_FOR_CLASSIFCATION = 20;
     private ITransactionalPreprocessingData preprocessingData;
 
@@ -63 +63 @@
     }
 
-    public List<double> GetVariableValues(string variableName) {
-      return preprocessingData.GetValues<double>(preprocessingData.GetColumnIndex(variableName)).ToList();
+    public IEnumerable<double> GetVariableValues(string variableName) {
+      return preprocessingData.GetValues<double>(preprocessingData.GetColumnIndex(variableName));
     }
 
@@ -74 +74 @@
         if (preprocessingData.IsType<double>(preprocessingData.GetColumnIndex(variableName)))
           doubleVariableNames.Add(variableName);
+      }
+
+      return doubleVariableNames;
+    }
+
+    public IEnumerable<string> GetVariableNamesForHistogramClassification() {
+      List<string> doubleVariableNames = new List<string>();
+
+      //only return variable names from type double
+      foreach (string variableName in preprocessingData.VariableNames)
+      {
+        int columnIndex           = preprocessingData.GetColumnIndex(variableName);
+        bool isDouble             = preprocessingData.IsType<double>(columnIndex);
+        double distinctValueCount = preprocessingData.GetValues<double>(columnIndex).GroupBy(x => x).Count();
+        bool distinctValuesOk     = distinctValueCount <= MAX_DISTINCT_VALUES_FOR_CLASSIFCATION;
+
+        if (isDouble && distinctValuesOk)
+              doubleVariableNames.Add(variableName);
       }
 
@@ -146 +164 @@
       
 
-      List<double> xValues = GetVariableValues(variableNameX);
-      List<double> yValues = GetVariableValues(variableNameY);
+      List<double> xValues = GetVariableValues(variableNameX).ToList();
+      List<double> yValues = GetVariableValues(variableNameY).ToList();
 
       List<Point2D<double>> points = new List<Point2D<double>>();