Weight Functions.Get Huber Weight Vector From Errors Method
Definition
Namespace: Numerics.NET.Curves
Assembly: Numerics.NET (in Numerics.NET.dll) Version: 9.0.3
Assembly: Numerics.NET (in Numerics.NET.dll) Version: 9.0.3
Overload List
Get | Returns a weight vector corresponding to the supplied error values using Huber's M-estimator. |
Get | Returns a weight vector corresponding to the supplied error values using Huber's M-estimator. |
GetHuberWeightVectorFromErrors(Vector<Double>)
Returns a weight vector corresponding to the supplied error values using Huber's M-estimator.
public static Vector<double> GetHuberWeightVectorFromErrors(
Vector<double> errors
)
Parameters
Return Value
Vector<Double>A vector containing the weights corresponding to the specified error values.
Remarks
This method uses Huber's M-estimator. The tuning constant is 1.345 times the standard deviation of the errors.
GetHuberWeightVectorFromErrors(Vector<Double>, Double)
Returns a weight vector corresponding to the supplied error values using Huber's M-estimator.
public static Vector<double> GetHuberWeightVectorFromErrors(
Vector<double> errors,
double tuningConstant
)
Parameters
- errors Vector<Double>
- A vector containing the errors for each observation.
- tuningConstant Double
- A tuning constant proportional to the largest error that can have the maximum weight.
Return Value
Vector<Double>A vector containing the weights corresponding to the specified error values.