refining log to ]0; +Inf[

[gro gro]

Hi there,

I'm currently running into trouble using leasqr function. In fact i'm making a non linear regression on the p vector using the function : 
y = (1/p(2))*(log((x-p(3))/EQF) - p(1));

as p(3) varies trought the leasqr algorithm and may produce negative values (x-p(3)<0), i may recover complex values.

I first decide to "redefine" a function log as follow:
function y=monlog(x)
if(x>0 && imag(x) == 0)
    y = log(x);
else
    y = Inf;
end

but the leasqr() function can't handle Inf values, so i'm trying to make the leasqr function to "forget" negative value by :
else
     y = -100;
 
using this it works "well", but i might now run into singular matrices and recover ugly covariance matrices (covp) from the function.

So i wonder...
Is there a way to *correctly* "redefine" this function under it's primal range [0; +Inf[ for leasqr to handle this
 correctly ?

Thanks for ideas and paths you may propose.
regards,
Gian

---

Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail