Changeset 15249 for stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions

Timestamp:

07/15/17 10:29:40 (7 years ago)

Author:

gkronber

Message:

#2699,#2700
merged r14862, r14863, r14911, r14936, r15156, r15157, r15158, r15164, r15169, r15207:15209, r15225, r15227, r15234, r15248 from trunk to stable

Location:

stable

Files:

• . (modified) (1 prop)
• HeuristicLab.Algorithms.DataAnalysis (modified) (1 prop)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression (copied) (copied from trunk/sources/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/CicularKernel.cs (modified) (2 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/GaussianKernel.cs (modified) (2 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/IKernel.cs (modified) (1 diff)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/InverseMultiquadraticKernel.cs (modified) (2 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/KernelBase.cs (modified) (4 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/MultiquadraticKernel.cs (modified) (3 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/PolysplineKernel.cs (modified) (3 diffs)
• HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/ThinPlatePolysplineKernel.cs (modified) (3 diffs)

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/CicularKernel.cs

@@ -25 +25 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   [Item("CircularKernel", "A circular kernel function 2*pi*(acos(-d)-d*(1-d²)^(0.5)) where n = ||x-c|| and d = n/beta \n  As described in http://crsouza.com/2010/03/17/kernel-functions-for-machine-learning-applications/")]
   public class CircularKernel : KernelBase {
-
-    #region HLConstructors & Boilerplate
     [StorableConstructor]
     protected CircularKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
+
     protected CircularKernel(CircularKernel original, Cloner cloner) : base(original, cloner) { }
-    public CircularKernel() {
-    }
+
+    public CircularKernel() { }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new CircularKernel(this, cloner);
     }
-    #endregion
 
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -53 +51 @@
     // 4*pi*n^3 / (beta^4 * sqrt(1-n^2/beta^2) 
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/GaussianKernel.cs

@@ -27 +27 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   [Item("GaussianKernel", "A kernel function that uses Gaussian function exp(-n²/beta²). As described in http://crsouza.com/2010/03/17/kernel-functions-for-machine-learning-applications/")]
   public class GaussianKernel : KernelBase {
-
-    #region HLConstructors & Boilerplate
     [StorableConstructor]
     protected GaussianKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
+
     protected GaussianKernel(GaussianKernel original, Cloner cloner) : base(original, cloner) { }
+
     public GaussianKernel() {
     }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new GaussianKernel(this, cloner);
     }
-    #endregion
 
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -54 +53 @@
     //2 * n²/b²* 1/b * exp(-n²/b²)
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/IKernel.cs

@@ -21 +21 @@
 
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+using System;
+
+namespace HeuristicLab.Algorithms.DataAnalysis {
   public interface IKernel : ICovarianceFunction {
     double? Beta { get; set; } // a kernel parameter
     IDistance Distance { get; set; } // the distance function to use
+
+    event EventHandler BetaChanged;
+    event EventHandler DistanceChanged;
   }
 }

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/InverseMultiquadraticKernel.cs

@@ -25 +25 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   [Item("InverseMultiquadraticKernel", "A kernel function that uses the inverse multi-quadratic function  1 / sqrt(1+||x-c||²/beta²). Similar to http://crsouza.com/2010/03/17/kernel-functions-for-machine-learning-applications/ with beta as a scaling factor.")]
   public class InverseMultiquadraticKernel : KernelBase {
+    private const double C = 1.0;
 
-    private const double C = 1.0;
-    #region HLConstructors & Boilerplate
     [StorableConstructor]
     protected InverseMultiquadraticKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
+
     protected InverseMultiquadraticKernel(InverseMultiquadraticKernel original, Cloner cloner) : base(original, cloner) { }
+
     public InverseMultiquadraticKernel() { }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new InverseMultiquadraticKernel(this, cloner);
     }
-    #endregion
 
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -52 +52 @@
     //n²/(b³(n²/b² + C)^1.5)
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/KernelBase.cs

@@ -28 +28 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   public abstract class KernelBase : ParameterizedNamedItem, IKernel {
 
-    #region Parameternames
     private const string DistanceParameterName = "Distance";
-    #endregion
-    #region Parameterproperties
-    public ValueParameter<IDistance> DistanceParameter {
-      get { return Parameters[DistanceParameterName] as ValueParameter<IDistance>; }
+
+    public IValueParameter<IDistance> DistanceParameter {
+      get { return (IValueParameter<IDistance>)Parameters[DistanceParameterName]; }
     }
 
     [Storable]
-    public double? Beta { get; set; }
-    #endregion
-    #region Properties
+    private double? beta;
+    public double? Beta {
+      get { return beta; }
+      set {
+        if (value != beta) {
+          beta = value;
+          RaiseBetaChanged();
+        }
+      }
+    }
+
     public IDistance Distance {
       get { return DistanceParameter.Value; }
-      set { DistanceParameter.Value = value; }
+      set {
+        if (DistanceParameter.Value != value) {
+          DistanceParameter.Value = value;
+        }
+      }
     }
-
-    #endregion
 
     [StorableConstructor]
     protected KernelBase(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
 
     protected KernelBase(KernelBase original, Cloner cloner)
       : base(original, cloner) {
-      Beta = original.Beta;
+      beta = original.beta;
+      RegisterEvents();
     }
 
@@ -64 +71 @@
       Parameters.Add(new ValueParameter<IDistance>(DistanceParameterName, "The distance function used for kernel calculation"));
       DistanceParameter.Value = new EuclideanDistance();
+      RegisterEvents();
+    }
+
+    [StorableHook(HookType.AfterDeserialization)]
+    private void AfterDeserialization() {
+      RegisterEvents();
+    }
+
+    private void RegisterEvents() {
+      DistanceParameter.ValueChanged += (sender, args) => RaiseDistanceChanged();
     }
 
@@ -82 +99 @@
     public ParameterizedCovarianceFunction GetParameterizedCovarianceFunction(double[] p, int[] columnIndices) {
       if (p.Length != GetNumberOfParameters(columnIndices.Length)) throw new ArgumentException("Illegal parametrization");
-      var myClone = (KernelBase)Clone(new Cloner());
+      var myClone = (KernelBase)Clone();
       myClone.SetParameter(p);
       var cov = new ParameterizedCovarianceFunction {
@@ -101 +118 @@
       return dist.Get(r1, r2);
     }
+
+    #region events
+    public event EventHandler BetaChanged;
+    public event EventHandler DistanceChanged;
+
+    protected void RaiseBetaChanged() {
+      var handler = BetaChanged;
+      if (handler != null) handler(this, EventArgs.Empty);
+    }
+
+    protected void RaiseDistanceChanged() {
+      var handler = DistanceChanged;
+      if (handler != null) handler(this, EventArgs.Empty);
+    }
+    #endregion
   }
 }

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/MultiquadraticKernel.cs

@@ -25 +25 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   // conditionally positive definite. (need to add polynomials) see http://num.math.uni-goettingen.de/schaback/teaching/sc.pdf 
@@ -32 +32 @@
 
     private const double C = 1.0;
-    #region HLConstructors & Boilerplate
+
     [StorableConstructor]
     protected MultiquadraticKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
-    protected MultiquadraticKernel(MultiquadraticKernel original, Cloner cloner)
-                : base(original, cloner) { }
 
-    public MultiquadraticKernel() {
-    }
+    protected MultiquadraticKernel(MultiquadraticKernel original, Cloner cloner) : base(original, cloner) { }
+
+    public MultiquadraticKernel() { }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new MultiquadraticKernel(this, cloner);
     }
-    #endregion
+
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -55 +54 @@
     //-n²/(d³*sqrt(C+n²/d²))
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/PolysplineKernel.cs

@@ -27 +27 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   // conditionally positive definite. (need to add polynomials) see http://num.math.uni-goettingen.de/schaback/teaching/sc.pdf 
@@ -33 +33 @@
   public class PolysplineKernel : KernelBase {
 
-    #region Parameternames
     private const string DegreeParameterName = "Degree";
-    #endregion
-    #region Parameterproperties
+
     public IFixedValueParameter<DoubleValue> DegreeParameter {
-      get { return Parameters[DegreeParameterName] as IFixedValueParameter<DoubleValue>; }
+      get { return (IFixedValueParameter<DoubleValue>)Parameters[DegreeParameterName]; }
     }
-    #endregion
-    #region Properties
+
     public DoubleValue Degree {
       get { return DegreeParameter.Value; }
     }
-    #endregion
 
-    #region HLConstructors & Boilerplate
     [StorableConstructor]
     protected PolysplineKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
+
     protected PolysplineKernel(PolysplineKernel original, Cloner cloner) : base(original, cloner) { }
+
     public PolysplineKernel() {
       Parameters.Add(new FixedValueParameter<DoubleValue>(DegreeParameterName, "The degree of the kernel. Needs to be greater than zero.", new DoubleValue(1.0)));
     }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new PolysplineKernel(this, cloner);
     }
-    #endregion
 
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -70 +66 @@
     //-degree/beta * (norm/beta)^degree
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;

stable/HeuristicLab.Algorithms.DataAnalysis/3.4/KernelRidgeRegression/KernelFunctions/ThinPlatePolysplineKernel.cs

@@ -27 +27 @@
 using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
 
-namespace HeuristicLab.Algorithms.DataAnalysis.KernelRidgeRegression {
+namespace HeuristicLab.Algorithms.DataAnalysis {
   [StorableClass]
   // conditionally positive definite. (need to add polynomials) see http://num.math.uni-goettingen.de/schaback/teaching/sc.pdf 
@@ -33 +33 @@
   public class ThinPlatePolysplineKernel : KernelBase {
 
-    #region Parameternames
     private const string DegreeParameterName = "Degree";
-    #endregion
-    #region Parameterproperties
+
     public IFixedValueParameter<DoubleValue> DegreeParameter {
-      get { return Parameters[DegreeParameterName] as IFixedValueParameter<DoubleValue>; }
+      get { return (IFixedValueParameter<DoubleValue>)Parameters[DegreeParameterName]; }
     }
-    #endregion
-    #region Properties
     public DoubleValue Degree {
       get { return DegreeParameter.Value; }
     }
-    #endregion
 
-    #region HLConstructors & Boilerplate
     [StorableConstructor]
     protected ThinPlatePolysplineKernel(bool deserializing) : base(deserializing) { }
-    [StorableHook(HookType.AfterDeserialization)]
-    private void AfterDeserialization() { }
+
     protected ThinPlatePolysplineKernel(ThinPlatePolysplineKernel original, Cloner cloner) : base(original, cloner) { }
+
     public ThinPlatePolysplineKernel() {
       Parameters.Add(new FixedValueParameter<DoubleValue>(DegreeParameterName, "The degree of the kernel. Needs to be greater than zero.", new DoubleValue(2.0)));
     }
+
     public override IDeepCloneable Clone(Cloner cloner) {
       return new ThinPlatePolysplineKernel(this, cloner);
     }
-    #endregion
 
     protected override double Get(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;
@@ -71 +66 @@
     // (Degree/beta) * (norm/beta)^Degree * log(norm/beta) 
     protected override double GetGradient(double norm) {
+      if (Beta == null) throw new InvalidOperationException("Can not calculate kernel distance gradient while Beta is null");
       var beta = Beta.Value;
       if (Math.Abs(beta) < double.Epsilon) return double.NaN;