I am trying to fit a 2d histogram using the following user-defined
function:
Double_t fitf(Double_t *x ,Double_t *par) {
Double_t k = x[0]+x[1];
Double_t q = x[0]-x[1];
Double_t kPower = 2.5;
Double_t qPower = 2.5;
Double_t fitval = par[0]+par[1]*pow(k,kPower)+par[2]*pow(q,qPower);
return fitval;
}
I have no problems at all as long as kPower and qPower are integers.
However, when (as above) the powers I specify for k and q are non-integers
the fit fails:
FCN=NaN FROM MIGRAD STATUS=CALL LIMIT 522 CALLS 523 TOTAL
EDM=NaN STRATEGY= 1 NO ERROR MATRIX
EXT PARAMETER CURRENT GUESS STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 Offset NaN 3.00000e-01 NaN NaN
2 kCoeff NaN 4.00000e-02 NaN NaN
3 qCoeff NaN 4.00000e-02 NaN NaN
I suspect it is a problem in calculating the derivative, but it seems that
it ought to be able to sucessfully calculate such a simple derivative.
Does anyone know what the problem might be?
---------------------------------------------------------------------
Jeffrey G Reid | jgreid@u.washington.edu | UW Nuclear Physics Lab
=====================================================================
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:17 MET