visualdatatools.com Forum Index visualdatatools.com
Discussion for DataTank and DataGraph
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Power Law and Exponential Decay

 
Post new topic   Reply to topic    visualdatatools.com Forum Index -> Request Fit Function
View previous topic :: View next topic  
Author Message
Jutho



Joined: 15 Aug 2007
Posts: 5

PostPosted: Tue Apr 14, 2009 10:19 am    Post subject: Power Law and Exponential Decay Reply with quote

Hi David,

two fitting functions I would like to see implemented are the exponential decay to a constant

y=a+b*exp(-c*x)

and the power law decay

y=a+b*x^(-c)

If possible, some statistical information about the uncertainty on a (and to a lesser extent c, for some applications) might be nice too. These functions are used a lot in finite size-scaling etc. Of course one can try to do these fits using the arbitrary function approach, but the results of that seem to be a bit disappointing to me. There's a strong dependence on the initial parameter values and the algorithm has a lot of trouble of finding an optimal fit. With the trial version of ProFit of QuantumSoft, I obtained very satisfactory results, but I don't want to buy that software if you could easily implement a better algorithm for fitting to these functions.

Sincerely,

Jutho
Back to top
View user's profile Send private message
David
Site Admin


Joined: 25 Nov 2006
Posts: 1940
Location: Chapel Hill, NC

PostPosted: Tue Apr 14, 2009 10:45 pm    Post subject: Reply with quote

I will harden then minimization routine.

David
Back to top
View user's profile Send private message Send e-mail
David
Site Admin


Joined: 25 Nov 2006
Posts: 1940
Location: Chapel Hill, NC

PostPosted: Sun May 31, 2009 1:37 am    Post subject: Reply with quote

Took some time, but it should be significantly better now. I also tweaked the UI a little bit. For some cases, and exponential fits often fit into this category the minima is very sensitive and the method can go into a funk and produce coefficients that are 10^9. But tweaking the initial value can help with that.

This will be part of 2.0.1, which will be released soon, and I just uploaded the beta.

Let me know how this stacks up against ProFit.

David
Back to top
View user's profile Send private message Send e-mail
Display posts from previous:   
Post new topic   Reply to topic    visualdatatools.com Forum Index -> Request Fit Function All times are GMT - 3 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group