LeastSquaresOptimizer.UseSparseJacobian Property

public bool UseSparseJacobian { get; set; }

Public Property UseSparseJacobian As Boolean
	Get
	Set

public:
property bool UseSparseJacobian {
	bool get ();
	void set (bool value);
}

member UseSparseJacobian : bool with get, set

Property Value

Set this property before calling FindMinimum() to avoid allocating a dense Jacobian at the start. Sparse matrices save memory when most derivatives are zero. Returning a sparse matrix from JacobianFunction automatically selects an iterative method for computing each optimization step.

The optimizer does not automatically identify zero derivatives when approximating the Jacobian numerically. Your Jacobian function must overwrite all entries that change, including entries that become zero. Requesting VarianceCovarianceMatrix still requires dense storage.

Reference