LeastSquaresOptimizer.JacobianVectorProduct Property

public Action<Vector<double>, Vector<double>, Vector<double>>? JacobianVectorProduct { get; set; }

Public Property JacobianVectorProduct As Action(Of Vector(Of Double), Vector(Of Double), Vector(Of Double))
	Get
	Set

public:
property Action<Vector<double>^, Vector<double>^, Vector<double>^>^ JacobianVectorProduct {
	Action<Vector<double>^, Vector<double>^, Vector<double>^>^ get ();
	void set (Action<Vector<double>^, Vector<double>^, Vector<double>^>^ value);
}

member JacobianVectorProduct : Action<Vector<float>, Vector<float>, Vector<float>> with get, set

Property Value

Supply this property together with JacobianTransposeVectorProduct. Both callbacks take precedence over JacobianFunction and select iterative steps without allocating a Jacobian matrix. The point and input are read-only. Fully overwrite the output with finite values and do not retain any of these buffers.

Products must describe the derivative of ObjectiveFunction at the supplied point. The optimizer uses unit column scaling in this mode. An explicit VarianceCovarianceMatrix request constructs a dense final Jacobian.

Reference