12.21.7. - The fitting equation
In the next part of the code you perform the actual evaluation of your fitting equation. The simple example shown is for fitting to a straight line.
double model(deriv_flag)
int deriv_flag;
{
double x, yfit;
x = X;
yfit = CONST + LINEAR * x;
if (deriv_flag) {
if (fCONST)
dCONST = 1;
if (fLINEAR)
dLINEAR = x;
}
return(yfit);
}
The argument
deriv_flag is nonzero when the routine
model() must calculate derivatives for the parameters currently being fit.
Of course, the routine only need calculate these derivatives if the
structure
initial indicates they are to be user-provided.
If
the derivatives are to be user-provided and
deriv_flag is set,
you still only calculate derivatives if the parameter is currently
being fit.
In this example, that is the case when
fCONST or
fLINEAR are nonzero.
Although you could use the definition
X for the current value
of the independent variable each time you refer to it, the code will
be more efficient if it refers to
X just once.
That is the
purpose of the line
x = X.