Changeset 13940

Timestamp:

06/28/16 12:53:54 (8 years ago)

Author:

gkronber

Message:

#745:

• added scatterplot of R² values over lambda instead of line chart,
• normalized coefficient values in coefficient path chart
• changed parameter lambda to LogLambda

File:

• branches/HeuristicLab.Algorithms.DataAnalysis.Glmnet/3.4/ElasticNetLinearRegression.cs (modified) (8 diffs)

branches/HeuristicLab.Algorithms.DataAnalysis.Glmnet/3.4/ElasticNetLinearRegression.cs

@@ -21 +21 @@
   public sealed class ElasticNetLinearRegression : FixedDataAnalysisAlgorithm<IRegressionProblem> {
     private const string PenalityParameterName = "Penality";
-    private const string LambdaParameterName = "Lambda";
+    private const string LogLambdaParameterName = "Log10(Lambda)";
     #region parameters
     public IFixedValueParameter<DoubleValue> PenalityParameter {
       get { return (IFixedValueParameter<DoubleValue>)Parameters[PenalityParameterName]; }
     }
-    public IValueParameter<DoubleValue> LambdaParameter {
-      get { return (IValueParameter<DoubleValue>)Parameters[LambdaParameterName]; }
+    public IValueParameter<DoubleValue> LogLambdaParameter {
+      get { return (IValueParameter<DoubleValue>)Parameters[LogLambdaParameterName]; }
     }
     #endregion
@@ -35 +35 @@
       set { PenalityParameter.Value.Value = value; }
     }
-    public DoubleValue Lambda {
-      get { return LambdaParameter.Value; }
-      set { LambdaParameter.Value = value; }
+    public DoubleValue LogLambda {
+      get { return LogLambdaParameter.Value; }
+      set { LogLambdaParameter.Value = value; }
     }
     #endregion
@@ -46 +46 @@
       : base(original, cloner) {
     }
-    public ElasticNetLinearRegression() : base() {
+    public ElasticNetLinearRegression()
+      : base() {
       Problem = new RegressionProblem();
       Parameters.Add(new FixedValueParameter<DoubleValue>(PenalityParameterName, "Penalty factor for balancing between ridge (0.0) and lasso (1.0) regression", new DoubleValue(0.5)));
-      Parameters.Add(new OptionalValueParameter<DoubleValue>(LambdaParameterName, "Optional: the value of lambda for which to calculate an elastic-net solution. lambda == null => calculate the whole path of all lambdas"));
+      Parameters.Add(new OptionalValueParameter<DoubleValue>(LogLambdaParameterName, "Optional: the value of lambda for which to calculate an elastic-net solution. lambda == null => calculate the whole path of all lambdas"));
     }
 
@@ -60 +61 @@
 
     protected override void Run() {
-      if (Lambda == null) {
+      if (LogLambda == null) {
         CreateSolutionPath();
       } else {
-        CreateSolution(Lambda.Value);
-      }
-    }
-
-    private void CreateSolution(double lambda) {
+        CreateSolution(LogLambda.Value);
+      }
+    }
+
+    private void CreateSolution(double logLambda) {
       double trainRsq;
       double testRsq;
-      var coeff = CreateElasticNetLinearRegressionSolution(Problem.ProblemData, Penality, lambda, out trainRsq, out testRsq);
+      var coeff = CreateElasticNetLinearRegressionSolution(Problem.ProblemData, Penality, Math.Pow(10, logLambda), out trainRsq, out testRsq);
       Results.Add(new Result("R² (train)", new DoubleValue(trainRsq)));
       Results.Add(new Result("R² (test)", new DoubleValue(testRsq)));
@@ -113 +114 @@
       RunElasticNetLinearRegression(Problem.ProblemData, Penality, out lambda, out trainRsq, out testRsq, out coeff, out intercept);
 
-      var coeffTable = new DataTable("Coefficient Paths", "The paths of coefficient values over different lambda values");
+      var coeffTable = new DataTable("Coefficient Paths", "The paths of standarized coefficient values over different lambda values");
       var nLambdas = lambda.Length;
       var nCoeff = coeff.GetLength(1);
@@ -120 +121 @@
       for (int i = 0; i < nCoeff; i++) {
         var coeffId = allowedVars[i];
-        var path = Enumerable.Range(0, nLambdas).Select(r => coeff[r, i]).ToArray();
+        double sigma = Problem.ProblemData.Dataset.GetDoubleValues(coeffId).StandardDeviation();
+        var path = Enumerable.Range(0, nLambdas).Select(r => coeff[r, i] * sigma).ToArray();
         dataRows[i] = new DataRow(coeffId, coeffId, path);
         coeffTable.Rows.Add(dataRows[i]);
@@ -127 +129 @@
       Results.Add(new Result(coeffTable.Name, coeffTable.Description, coeffTable));
 
-      var rsqTable = new DataTable("R-Squared", "Path of R² values over different lambda values");
-      rsqTable.Rows.Add(new DataRow("R² (train)", "Path of R² values over different lambda values", trainRsq));
-      rsqTable.Rows.Add(new DataRow("R² (test)", "Path of R² values over different lambda values", testRsq));
-      rsqTable.Rows.Add(new DataRow("Lambda", "The lambda values along the path", lambda));
-      rsqTable.Rows["Lambda"].VisualProperties.ChartType = DataRowVisualProperties.DataRowChartType.Points;
-      rsqTable.Rows["Lambda"].VisualProperties.SecondYAxis = true;
-      rsqTable.VisualProperties.SecondYAxisMinimumFixedValue = 1E-7;
-      rsqTable.VisualProperties.SecondYAxisLogScale = true;
-      Results.Add(new Result(rsqTable.Name, rsqTable.Description, rsqTable));
+      var rsqPlot = new ScatterPlot("R-Squared", "Path of R² values over different lambda values");
+      rsqPlot.VisualProperties.YAxisMaximumAuto = false;
+      rsqPlot.VisualProperties.YAxisMinimumAuto = false;
+      rsqPlot.VisualProperties.XAxisMaximumAuto = false;
+      rsqPlot.VisualProperties.XAxisMinimumAuto = false;
+
+      rsqPlot.VisualProperties.YAxisMinimumFixedValue = 0;
+      rsqPlot.VisualProperties.YAxisMaximumFixedValue = 1.0;
+      rsqPlot.VisualProperties.XAxisTitle = "Log10(Lambda)";
+      rsqPlot.VisualProperties.YAxisTitle = "R²";
+      rsqPlot.Rows.Add(new ScatterPlotDataRow("R² (train)", "Path of R² values over different lambda values", lambda.Zip(trainRsq, (l, r) => new Point2D<double>(Math.Log10(l), r))));
+      rsqPlot.Rows.Add(new ScatterPlotDataRow("R² (test)", "Path of R² values over different lambda values", lambda.Zip(testRsq, (l, r) => new Point2D<double>(Math.Log10(l), r))));
+      if (lambda.Length > 2) {
+        rsqPlot.VisualProperties.XAxisMinimumFixedValue = Math.Floor(Math.Log10(lambda.Last()));
+        rsqPlot.VisualProperties.XAxisMaximumFixedValue = Math.Ceiling(Math.Log10(lambda.Skip(1).First()));
+      }
+      rsqPlot.Rows["R² (train)"].VisualProperties.PointSize = 5;
+      rsqPlot.Rows["R² (test)"].VisualProperties.PointSize = 5;
+
+      Results.Add(new Result(rsqPlot.Name, rsqPlot.Description, rsqPlot));
     }
 
@@ -227 +240 @@
       int nx = numVars;
       double thr = 1.0e-5; // default value as recommended in glmnet
-      int isd = 0; //  => regression on original predictor variables
+      int isd = 1; //  => regression on standardized predictor variables
       int intr = 1;  // => do include intercept in model
       int maxit = 100000; // default value as recommended in glmnet