Changeset 10815 for branches/DataPreprocessing

Timestamp:

05/07/14 14:23:28 (10 years ago)

Author:

tsteinre

Message:

• implemented StandardDeviationTransformation logic.

Location:

branches/DataPreprocessing/HeuristicLab.Problems.DataAnalysis.Transformations/3.3

Files:

• HeuristicLab.Problems.DataAnalysis.Transformations-3.3.csproj (modified) (1 diff)
• StandardDeviationTransformation.cs (moved) (moved from branches/DataPreprocessing/HeuristicLab.Problems.DataAnalysis.Transformations/3.3/DeviationTransformation.cs) (2 diffs)

branches/DataPreprocessing/HeuristicLab.Problems.DataAnalysis.Transformations/3.3/HeuristicLab.Problems.DataAnalysis.Transformations-3.3.csproj

@@ -48 +48 @@
   <ItemGroup>
     <Compile Include="CopyColumnTransformation.cs" />
-    <Compile Include="DeviationTransformation.cs" />
+    <Compile Include="StandardDeviationTransformation.cs" />
     <Compile Include="PowerTransformation.cs" />
     <Compile Include="ExponentialTransformation.cs" />

branches/DataPreprocessing/HeuristicLab.Problems.DataAnalysis.Transformations/3.3/StandardDeviationTransformation.cs

@@ -1 @@
-using System;
-using System.Collections.Generic;
+using System.Collections.Generic;
 using System.Linq;
 using HeuristicLab.Common;
@@ -9 +8 @@
 
 namespace HeuristicLab.Problems.DataAnalysis.Transformations {
-  [Item("DeviationTransformation", "Represents a linear Transformation using deviation Parameters")]
-  public class DeviationTransformation : LinearTransformation {
+  [Item("StandardDeviationTransformation", "f(x) = ((x - m_org) / s_org ) * s_tar + m_tar | Represents Transformation to unit standard deviation and additional linear transformation to a target Mean and Standard deviation")]
+  public class StandardDeviationTransformation : Transformation<double> {
+    protected const string OriginalMeanParameterName = "OriginalMean";
+    protected const string OriginalStandardDeviationParameterName = "OriginalStandardDeviation";
     protected const string MeanParameterName = "Mean";
-    protected const string VarianceParameterName = "Variance";
+    protected const string StandardDeviationParameterName = "StandardDeviation";
 
     #region Parameters
+    public IValueParameter<DoubleValue> OriginalMeanParameter {
+      get { return (IValueParameter<DoubleValue>)Parameters[OriginalMeanParameterName]; }
+    }
+    public IValueParameter<DoubleValue> OriginalStandardDeviationParameter {
+      get { return (IValueParameter<DoubleValue>)Parameters[OriginalStandardDeviationParameterName]; }
+    }
     public IValueParameter<DoubleValue> MeanParameter {
       get { return (IValueParameter<DoubleValue>)Parameters[MeanParameterName]; }
     }
-    public IValueParameter<DoubleValue> VarianceParameter {
-      get { return (IValueParameter<DoubleValue>)Parameters[VarianceParameterName]; }
+    public IValueParameter<DoubleValue> StandardDeviationParameter {
+      get { return (IValueParameter<DoubleValue>)Parameters[StandardDeviationParameterName]; }
     }
     #endregion
 
     #region properties
+    public double OriginalMean {
+      get { return OriginalMeanParameter.Value.Value; }
+      set { OriginalMeanParameter.Value.Value = value; }
+    }
+    public double OriginalStandardDeviation {
+      get { return OriginalStandardDeviationParameter.Value.Value; }
+      set { OriginalStandardDeviationParameter.Value.Value = value; }
+    }
     public double Mean {
       get { return MeanParameter.Value.Value; }
     }
-    public double Variance {
-      get { return VarianceParameter.Value.Value; }
+    public double StandardDeviation {
+      get { return StandardDeviationParameter.Value.Value; }
     }
     #endregion
 
     [StorableConstructor]
-    protected DeviationTransformation(bool deserializing) : base(deserializing) { }
-    protected DeviationTransformation(Transformation<double> original, Cloner cloner)
+    protected StandardDeviationTransformation(bool deserializing) : base(deserializing) { }
+    protected StandardDeviationTransformation(Transformation<double> original, Cloner cloner)
       : base(original, cloner) {
     }
-    public DeviationTransformation(IEnumerable<string> allowedColumns)
+    public StandardDeviationTransformation(IEnumerable<string> allowedColumns)
       : base(allowedColumns) {
-      MultiplierParameter.Hidden = true;
-      AddendParameter.Hidden = true;
-      Parameters.Add(new ValueParameter<DoubleValue>(MeanParameterName, "Mean value for the target deviation.", new DoubleValue(0.0)));
-      Parameters.Add(new ValueParameter<DoubleValue>(VarianceParameterName, "Variance for the target deviation.", new DoubleValue(1.0)));
+      Parameters.Add(new ValueParameter<DoubleValue>(OriginalMeanParameterName, "m_org | Mean value of the original data's deviation.", new DoubleValue()));
+      Parameters.Add(new ValueParameter<DoubleValue>(OriginalStandardDeviationParameterName, "s_org | Standard deviation of the original data.", new DoubleValue()));
+      OriginalMeanParameter.Hidden = true;
+      OriginalStandardDeviationParameter.Hidden = true;
+      Parameters.Add(new ValueParameter<DoubleValue>(MeanParameterName, "m_tar | Mean value for the target deviation.", new DoubleValue(0.0)));
+      Parameters.Add(new ValueParameter<DoubleValue>(StandardDeviationParameterName, "s_tar | Standard deviation for the target data.", new DoubleValue(1.0)));
     }
 
     public override IDeepCloneable Clone(Cloner cloner) {
-      return new DeviationTransformation(this, cloner);
+      return new StandardDeviationTransformation(this, cloner);
     }
 
+    // http://en.wikipedia.org/wiki/Standard_deviation
+    // http://www.statistics4u.info/fundstat_germ/ee_ztransform.html
+    // https://www.uni-due.de/~bm0061/vorl12.pdf p5
     public override IEnumerable<double> Apply(IEnumerable<double> data) {
       ConfigureParameters(data);
-      return base.Apply(data);
+      if (OriginalStandardDeviation == 0.0) {
+        foreach (var e in data) {
+          yield return e;
+        }
+        yield break;
+      }
+
+      foreach (var e in data) {
+        double unitNormalDistributedValue = (e - OriginalMean) / OriginalStandardDeviation;
+        yield return unitNormalDistributedValue * StandardDeviation + Mean;
+      }
     }
 
     public override bool Check(IEnumerable<double> data, out string errorMsg) {
       ConfigureParameters(data);
-      return base.Check(data, out errorMsg);
+      errorMsg = "";
+      if (OriginalStandardDeviation == 0.0) {
+        errorMsg = "Standard deviaton for the original data is 0.0";
+        return false;
+      } else {
+        return true;
+      }
     }
 
     protected void ConfigureParameters(IEnumerable<double> data) {
-      double standardDeviation = data.StandardDeviation();
-      double middle = data.Average();
-      // TODO:
-      // http://en.wikipedia.org/wiki/Standard_deviation
-      // http://www.statistics4u.info/fundstat_germ/ee_ztransform.html
-      // https://www.uni-due.de/~bm0061/vorl12.pdf p5
-      throw new NotImplementedException();
+      OriginalStandardDeviation = data.StandardDeviation();
+      OriginalMean = data.Average();
     }
-
   }
 }