Hi,
the best fit I could do is the following in the plot by removing
your limits.
The limits are causing the function to move to values in more
disagreement with the data.
P.S. In your case when you are using a standard least quare function
from an histogram is much simpler to call the TH1::Fit method passing
your
model function (called "nm(x,p)" in your code).
You can also use the FitPanel (opend from the "Tool" menu of TCanvas
or by right clicking on the data) to play inetractivly with limits
and initial parameters.
Lorenzo
On Aug 16, 2009, at 9:55 AM, Anil Singh wrote:
> Dear Virender,
>
> I would suggest you to try following and see if the results change:
>
> 1_ Fix one of the parameters(probably the one that stays at lower
> end), limit other two and do the fit.
> 2_Release the parameter, and fix the other one and again do the fit.
>
> A Singh
>
>
>
> ----- Original Message -----
> From: vicky mangat <mangat_phy_at_yahoo.co.in>
> Date: Friday, August 14, 2009 12:27 pm
> Subject: [ROOT] Problem in Minuit
> To: roottalk_at_root.cern.ch
>
>
>>
>> Hi
>> I was try to fit a function using Minuit. I am able to write
>> program for this and it is excuting very well. But i have found few
>> problem in this.
>> 1. Quality of fit is poor. I have a attached root file of plot you
>> can see it.
>> 2. Secondly I have found that value of my first parameter always
>> fix at lower limit applied by me.
>> 0.1 3.4412897 2.530768 2.5043414
>> 329.21223
>> chi2: 329.21223
>> The Chi2/ndf is: 8.8976279
>> Here i have given limit 0.1 to 6 on first parameter.
>> Now if i change this limit say 2 to 6 then parameters are
>> 2 2.2054358 2.530847 2.5044359
>> 329.51952
>> chi2: 329.51952
>> The Chi2/ndf is: 8.905933
>>
>> 3. My chi2 per degree of freedom is found to be high . Can i
>> imporve it.
>> I have attached all program files and data file so that i
>> you want you can run in on ur own pc.
>> Thankx.
>>
>>
>> Varinderjit Singh
>>
>> Research Scholar
>>
>> Dept. of Physics
>>
>> Panjab University
>>
>> Chandigarh
>>
>> 09914100789
>>
>> Yahoo! recommends that you upgrade to the new and safer
>> Internet Explorer 8.
>>
>>
>> Looking for local information? Find it on Yahoo! Local http://in.local.yahoo.com/
>
This archive was generated by hypermail 2.2.0 : Wed Aug 19 2009 - 23:50:02 CEST