Changeset 5658 for branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4

Timestamp:

03/10/11 12:38:43 (13 years ago)

Author:

gkronber

Message:

#1418 implemented wrapper for LDA (linear discriminant analysis) implemented in alglib.

Location:

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4

Files:

• HeuristicLab.Algorithms.DataAnalysis-3.4.csproj (modified) (1 diff)
• Linear/AlglibUtil.cs (moved) (moved from branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/Linear/LinearRegressionUtil.cs) (1 diff)
• Linear/LinearDiscriminantAnalysis.cs (added)
• Linear/LinearRegression.cs (modified) (1 diff)
• kMeans/KMeansClustering.cs (modified) (1 diff)
• kMeans/KMeansClusteringModel.cs (modified) (1 diff)
• kMeans/KMeansClusteringUtil.cs (modified) (1 diff)

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/HeuristicLab.Algorithms.DataAnalysis-3.4.csproj

@@ -116 +116 @@
     <Compile Include="kMeans\KMeansClusteringModel.cs" />
     <Compile Include="kMeans\KMeansClusteringSolution.cs" />
-    <Compile Include="kMeans\KMeansClusteringUtil.cs" />
+    <Compile Include="kMeans\KMeansClusteringUtil.cs">
+      <SubType>Code</SubType>
+    </Compile>
+    <Compile Include="Linear\AlglibUtil.cs" />
+    <Compile Include="Linear\LinearDiscriminantAnalysis.cs" />
     <Compile Include="Linear\LinearRegression.cs">
-      <SubType>Code</SubType>
-    </Compile>
-    <Compile Include="Linear\LinearRegressionUtil.cs">
       <SubType>Code</SubType>
     </Compile>

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/Linear/AlglibUtil.cs

@@ -25 +25 @@
 
 namespace HeuristicLab.Algorithms.DataAnalysis {
-  public static class LinearRegressionUtil {
-    public static double[,] PrepareInputMatrix(Dataset dataset, string targetVariable, IEnumerable<string> allowedInputVariables, int start, int end) {
-      List<int> allowedRows = CalculateAllowedRows(dataset, targetVariable, allowedInputVariables, start, end);
+  public static class AlglibUtil {
+    public static double[,] PrepareInputMatrix(Dataset dataset, IEnumerable<string> variables, IEnumerable<int> rows) {
+      List<int> allowedRows = CalculateAllowedRows(dataset, variables, rows).ToList();
 
-      double[,] matrix = new double[allowedRows.Count, allowedInputVariables.Count() + 1];
+      double[,] matrix = new double[allowedRows.Count, variables.Count()];
       for (int row = 0; row < allowedRows.Count; row++) {
         int col = 0;
-        foreach (string column in allowedInputVariables) {
+        foreach (string column in variables) {
           matrix[row, col] = dataset[column, row];
           col++;
         }
-        matrix[row, allowedInputVariables.Count()] = dataset[targetVariable, row];
       }
       return matrix;
     }
 
-    private static List<int> CalculateAllowedRows(Dataset dataset, string targetVariable, IEnumerable<string> allowedInputVariables, int start, int end) {
-      List<int> allowedRows = new List<int>();
-      bool add = false;
-
-      for (int row = start; row < end; row++) {
-        add = true;
-        foreach (string column in allowedInputVariables) {
-          double value = dataset[column, row];
-          if (double.IsInfinity(value) ||
-            double.IsNaN(value))
-            add = false;
-        }
-        if (double.IsNaN(dataset[targetVariable, row]))
-          add = false;
-        if (add)
-          allowedRows.Add(row);
-        add = true;
-      }
-      return allowedRows;
+    private static IEnumerable<int> CalculateAllowedRows(Dataset dataset, IEnumerable<string> variables, IEnumerable<int> rows) {
+      // return only rows that contain no infinity or NaN values
+      return from row in rows
+             where (from variable in variables
+                    let x = dataset[variable, row]
+                    where double.IsInfinity(x) || double.IsNaN(x)
+                    select 1)
+                    .Any() == false
+             select row;
     }
   }

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/Linear/LinearRegression.cs

@@ -75 +75 @@
       int samplesStart = problemData.TrainingPartitionStart.Value;
       int samplesEnd = problemData.TrainingPartitionEnd.Value;
-
-      double[,] inputMatrix = LinearRegressionUtil.PrepareInputMatrix(dataset, targetVariable, allowedInputVariables, samplesStart, samplesEnd);
+      IEnumerable<int> rows = Enumerable.Range(samplesStart, samplesEnd - samplesStart);
+      double[,] inputMatrix = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables.Concat(new string[] { targetVariable }), rows);
 
       alglib.linearmodel lm = new alglib.linearmodel();

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/kMeans/KMeansClustering.cs

@@ -94 +94 @@
       double[,] centers;
       int[] xyc;
-      double[,] inputMatrix = KMeansClusteringUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
+      double[,] inputMatrix = AlglibUtil.PrepareInputMatrix(dataset, allowedInputVariables, rows);
       alglib.kmeansgenerate(inputMatrix, inputMatrix.GetLength(0), inputMatrix.GetLength(1), k, restarts + 1, out info, out centers, out xyc);
       if (info != 1) throw new ArgumentException("Error in calculation of k-Means clustering solution");

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/kMeans/KMeansClusteringModel.cs

@@ -64 +64 @@
     public KMeansClusteringModel(double[,] centers, IEnumerable<string> allowedInputVariables)
       : base() {
+      this.name = ItemName;
+      this.description = ItemDescription;
       // disect center matrix into list of double[]
       // centers are given as double matrix where number of rows = dimensions and number of columns = clusters

branches/DataAnalysis Refactoring/HeuristicLab.Algorithms.DataAnalysis/3.4/kMeans/KMeansClusteringUtil.cs

@@ -27 +27 @@
 namespace HeuristicLab.Algorithms.DataAnalysis {
   public static class KMeansClusteringUtil {
-    public static double[,] PrepareInputMatrix(Dataset dataset, IEnumerable<string> allowedInputVariables, IEnumerable<int> rows) {
-      List<int> allowedRows = CalculateAllowedRows(dataset, allowedInputVariables, rows).ToList();
-
-      double[,] matrix = new double[allowedRows.Count, allowedInputVariables.Count()];
-      for (int row = 0; row < allowedRows.Count; row++) {
-        int col = 0;
-        foreach (string column in allowedInputVariables) {
-          matrix[row, col] = dataset[column, row];
-          col++;
-        }
-      }
-      return matrix;
-    }
-
-    private static IEnumerable<int> CalculateAllowedRows(Dataset dataset, IEnumerable<string> allowedInputVariables, IEnumerable<int> rows) {
-      // return only rows that contain no infinity or NaN values
-      return from row in rows
-             where (from inputVariable in allowedInputVariables
-                    let x = dataset[inputVariable, row]
-                    where double.IsInfinity(x) || double.IsNaN(x)
-                    select 1)
-                    .Any() == false
-             select row;
-    }
-
     public static IEnumerable<int> FindClosestCenters(IEnumerable<double[]> centers, Dataset dataset, IEnumerable<string> allowedInputVariables, IEnumerable<int> rows) {
       int nRows = rows.Count();