TMtxMultiNonLinReg::DeriveProcedure Property

__declspec(property(get = GetDeriveProcedure, put = SetDeriveProcedure)) TMultiDeriveProc DeriveProcedure;

Defines derivatives of regression function.

Read: TMultiDeriveProc GetDeriveProcedure();
Write: void SetDeriveProcedure(const TMultiDeriveProc Value);
Remarks:

Defines the procedure for calculating the derivatives of a regression function TMultiRegressFun with respect to the regression parameters TMtxMultiNonLinReg::B, evaluated at specific set of (TMtxMultiNonLinReg::X,TMtxMultiNonLinReg::Y) points.

If DeriveProcedure is not declared (for example if it's nil) TMtxMultiNonLinReg will use the approximate numeric derivative. In this case the gradient step size is defined by GradientStepSize (Optimization unit) global variable. Normally the optimal stepsize depends on seventh partial derivatives of the function. Since they (in this case) are not available, the initial value for GradientStepSize is Exp(Ln(EPS)/7)*0.25 (as suggested by Spellucci).

Declared in Dew::Stats::TMtxMultiNonLinReg · Dew.Stats/StatTools.h · Cross-compiler