You are here: Symbol Reference > Regress Namespace > Functions > Regress.MultiNumericDerive Function
Vectorized computation of the numerical derivative for the regress function.
procedure MultiNumericDerive(const RegressFun: TMultiRegressFun; const X: TVecList; const Y: TVec; const Pars: TVec; const Grad: TVecList);
Routine calculates the derivatives of the regression function RegressFun at (X,Y) by numerical differentiation. The results of differentation are returned in Grad vectors, one vector for each parameter stored in "pars". The "X" stores the independent variables and "Y" the dependent variable. RegressFun is the function, which computes the value of the function of which we are looking for derivates.
The Length and Complex properties of Grad vector are NOT adjusted automatically. Grad.Count should typically be equal to Pars.Length and Grad[i].Length equal to Y.Length. The function returns derivates separately for each parameter and pair of (X0...Xi, Y).
Vectorized computation of the derivates can be 10-20x faster than non-vectorized computation.
|
Copyright (c) 1999-2025 by Dew Research. All rights reserved.
|
|
What do you think about this topic? Send feedback!
|
File