Migrad doesn't converge , also how do we calculate error matrix letting first12 parameters fixed and 13-21 parameters free?

Roofit and RooStats
1

Dear Rooters !
I wrote a code(minuit1.c) to minimize chisquare and calculate error matrix as well. I have 21 parameters and 1328 datapoints. I need to calculate error matrix with first 12 parameters fixed and letting 13-21 parameters free. I know that I have messed up a lot, but don’t know how to make those correct. Can anyone help me in it please?
I got different types of errors: Status:limit calls and finally STATUS FAILED. minuit1.c (10.4 KB)

2

Hi,
to run the code we need also your input data. Please add is as attachment or add a link that can be used for downloading the data.

3

Hi
Here is the data attachedallDataForMinuit.txt (55.0 KB)

Thanks
Pushpa

4

Also, to add here, all the parameters have to be free when minimizing chisquare, but to find error matrix, first 12 parameters have to be fixed and the last 9 have to be free to get 9 by 9 matrix.

5

It is called the “partial covariance matrix”.

6

It looks to me that the derivative of te first parameter is zero. So you can fix itto remove from the minimization.
Then if you are interested on the correlations of 9x9 parameters get the full matrix (but make sure to run Hesse successfully) and then you get the sub-matrix for the 9x9 parameters using as example the link posted before

7

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.