Set default values for regularization and outlier removal so user isn't forced to provide those parameters

Author: ooples

src/Models/RegularizationOptions.cs

@@ -4,5 +4,5 @@ public class RegularizationOptions
 {
     public RegularizationType Type { get; set; } = RegularizationType.None;
     public double Strength { get; set; } = 0.0;
-    public double L1Ratio { get; set; } = 0.5; // Only used for ElasticNet
+    public double L1Ratio { get; set; } = 0.5;
 }

src/OutlierRemoval/IQROutlierRemoval.cs

@@ -5,10 +5,10 @@ public class IQROutlierRemoval<T> : IOutlierRemoval<T>
     private readonly T _iqrMultiplier;
     private readonly INumericOperations<T> _numOps;
 
-    public IQROutlierRemoval(T iqrMultiplier)
+    public IQROutlierRemoval(T? iqrMultiplier = default)
     {
-        _iqrMultiplier = iqrMultiplier;
         _numOps = MathHelper.GetNumericOperations<T>();
+        _iqrMultiplier = iqrMultiplier ?? GetDefaultMultiplier();
     }
 
     public (Matrix<T> CleanedInputs, Vector<T> CleanedOutputs) RemoveOutliers(Matrix<T> inputs, Vector<T> outputs)
@@ -45,4 +45,9 @@ public IQROutlierRemoval(T iqrMultiplier)
 
         return (new Matrix<T>(cleanedInputs, _numOps), new Vector<T>(cleanedOutputs, _numOps));
     }
+
+    private T GetDefaultMultiplier()
+    {
+        return _numOps.FromDouble(1.5);
+    }
 }

src/OutlierRemoval/MADOutlierRemoval.cs

@@ -5,10 +5,10 @@ public class MADOutlierRemoval<T> : IOutlierRemoval<T>
     private readonly T _threshold;
     private readonly INumericOperations<T> _numOps;
 
-    public MADOutlierRemoval(T threshold)
+    public MADOutlierRemoval(T? threshold = default)
     {
-        _threshold = threshold;
         _numOps = MathHelper.GetNumericOperations<T>();
+        _threshold = threshold ?? GetDefaultThreshold();
     }
 
     public (Matrix<T> CleanedInputs, Vector<T> CleanedOutputs) RemoveOutliers(Matrix<T> inputs, Vector<T> outputs)
@@ -43,4 +43,9 @@ public MADOutlierRemoval(T threshold)
 
         return (new Matrix<T>(cleanedInputs, _numOps), new Vector<T>(cleanedOutputs, _numOps));
     }
+
+    private T GetDefaultThreshold()
+    {
+        return _numOps.FromDouble(3.5);
+    }
 }

src/OutlierRemoval/ThresholdOutlierRemoval.cs

@@ -1,17 +1,14 @@
 namespace AiDotNet.OutlierRemoval;
 
-/// <summary>
-/// Removes outliers from the data using the threshold method. This method is not recommended for data sets with less than 15 data points.
-/// </summary>
 public class ThresholdOutlierRemoval<T> : IOutlierRemoval<T>
 {
     private readonly T _threshold;
     private readonly INumericOperations<T> _numOps;
 
-    public ThresholdOutlierRemoval(T threshold)
+    public ThresholdOutlierRemoval(T? threshold = default)
     {
-        _threshold = threshold;
         _numOps = MathHelper.GetNumericOperations<T>();
+        _threshold = threshold ?? GetDefaultThreshold();
     }
 
     public (Matrix<T> CleanedInputs, Vector<T> CleanedOutputs) RemoveOutliers(Matrix<T> inputs, Vector<T> outputs)
@@ -42,4 +39,9 @@ public ThresholdOutlierRemoval(T threshold)
 
         return (new Matrix<T>(cleanedInputs, _numOps), new Vector<T>(cleanedOutputs, _numOps));
     }
+
+    private T GetDefaultThreshold()
+    {
+        return _numOps.FromDouble(3.0);
+    }
 }

src/OutlierRemoval/ZScoreOutlierRemoval.cs

@@ -5,10 +5,10 @@ public class ZScoreOutlierRemoval<T> : IOutlierRemoval<T>
     private readonly T _threshold;
     private readonly INumericOperations<T> _numOps;
 
-    public ZScoreOutlierRemoval(T threshold)
+    public ZScoreOutlierRemoval(T? threshold = default)
     {
-        _threshold = threshold;
         _numOps = MathHelper.GetNumericOperations<T>();
+        _threshold = threshold ?? FindDefaultThreshold();
     }
 
     public (Matrix<T> CleanedInputs, Vector<T> CleanedOutputs) RemoveOutliers(Matrix<T> inputs, Vector<T> outputs)
@@ -41,4 +41,9 @@ public ZScoreOutlierRemoval(T threshold)
 
         return (new Matrix<T>(cleanedInputs, _numOps), new Vector<T>(cleanedOutputs, _numOps));
     }
+
+    private T FindDefaultThreshold()
+    {
+        return _numOps.FromDouble(3.0);
+    }
 }

src/Regularization/ElasticRegularization.cs

@@ -5,10 +5,10 @@ public class ElasticNetRegularization<T> : IRegularization<T>
     private readonly INumericOperations<T> _numOps;
     private readonly RegularizationOptions _options;
 
-    public ElasticNetRegularization(INumericOperations<T> numOps, RegularizationOptions options)
+    public ElasticNetRegularization(RegularizationOptions? options = null)
     {
-        _numOps = numOps;
-        _options = options;
+        _numOps = MathHelper.GetNumericOperations<T>();
+        _options = options ?? new RegularizationOptions();
     }
 
     public Matrix<T> RegularizeMatrix(Matrix<T> matrix)

src/Regularization/L1Regularization.cs

@@ -5,15 +5,14 @@ public class L1Regularization<T> : IRegularization<T>
     private readonly INumericOperations<T> _numOps;
     private readonly RegularizationOptions _options;
 
-    public L1Regularization(INumericOperations<T> numOps, RegularizationOptions options)
+    public L1Regularization(RegularizationOptions? options = null)
     {
-        _numOps = numOps;
-        _options = options;
+        _numOps = MathHelper.GetNumericOperations<T>();
+        _options = options ?? new RegularizationOptions();
     }
 
     public Matrix<T> RegularizeMatrix(Matrix<T> matrix)
     {
-        // L1 regularization doesn't modify the matrix directly
         return matrix;
     }
 

src/Regularization/L2Regularization.cs

@@ -5,10 +5,10 @@ public class L2Regularization<T> : IRegularization<T>
     private readonly INumericOperations<T> _numOps;
     private readonly RegularizationOptions _options;
 
-    public L2Regularization(INumericOperations<T> numOps, RegularizationOptions options)
+    public L2Regularization(RegularizationOptions? options = null)
     {
-        _numOps = numOps;
-        _options = options;
+        _numOps = MathHelper.GetNumericOperations<T>();
+        _options = options ?? new RegularizationOptions();
     }
 
     public Matrix<T> RegularizeMatrix(Matrix<T> matrix)