Welcome Guest! To enable all features please Login. New Registrations are disabled.

Notification

Icon
Error

Login


Options
Go to last post Go to first unread
Offline Jean Giraud  
#1 Posted : 20 February 2017 23:33:28(UTC)
Jean Giraud

Rank: Guest

Groups: Registered
Joined: 04/07/2015(UTC)
Posts: 6,868
Canada

Was thanked: 980 time(s) in 808 post(s)
... another good example, observe Pearson correlation.

Jean

Genfit Rational INITIALIZE.sm (64kb) downloaded 41 time(s).
thanks 1 user thanked Jean Giraud for this useful post.
on 19/03/2017(UTC)

Wanna join the discussion?! Login to your SMath Studio Forum forum account. New Registrations are disabled.

Offline djlischer  
#2 Posted : 19 March 2017 18:31:15(UTC)
djlischer

Rank: Newbie

Groups: Registered
Joined: 06/03/2017(UTC)
Posts: 2
United States
Location: Acton, MA

Jean,

Thank you for your file on Genfit curving fitting with a rational function. I'm new to SMath and I learned a lot from this file! I've made a few small changes in this new Rev2 version that you may or may not find useful.

The biggest change was that I used the methodology mentioned here to initialize all the fit terms rather than just the numerator terms. If there are 'k' terms in the fit equation, you simply enter 'k' data points that you want the initization to fit exactly. The program then solves linearly for the required fit parameters. This seems to work pretty well, but I found the results can be sensitive to the initially fitted data points.

Also, I changed how the rational equation is entered. Now you can just enter they desired polynomial orders for the numerator and the denominator and the program develops the equation itself. This makes it easier to experiment with different fits since you don't have to retype the equation each time. I also added a function for Pearson's correlation coefficient to make it easier to look at the correlation of various fits.

Regards,
Jeff

Genfit Rational INITIALIZE_Rev2.sm (92kb) downloaded 28 time(s).
Offline Jean Giraud  
#3 Posted : 21 March 2017 04:05:02(UTC)
Jean Giraud

Rank: Guest

Groups: Registered
Joined: 04/07/2015(UTC)
Posts: 6,868
Canada

Was thanked: 980 time(s) in 808 post(s)
Jeff,

Thanks for your attentive input. I'm familiar to setup the model num/den.
Your proposal is what we have developed in Mathcad "GenfitMatrix".
In this application, your initials is just great, further refined "Minimize".
Differently than the PT100 [Platinum Resistance Thermomether], I found no
relationship between the TC 'J' data set and an eventual XFR [Transfer Function].
Thus: found no model function. True: initializing is capital and there is no rule.
Fitting data set is an art [F.B. Hildebrand].

Oh ! NIST muffed Han, Kirby. These two have a model function [attached].
Are you familiar with Mathcad ? My Mathcad 11 still working.
Mathcad 11 "GenfitMatrix" ranks first in my tool box.
If you are interested: can mail to you in case Smath forum couldn't take.

Thanks, most appreciated.

Genfit Kirby_2 [Transmute].sm (34kb) downloaded 21 time(s).

Genfit Kirby_2 [Reduced].sm (38kb) downloaded 21 time(s).

Genfit Rational INITIALIZE_Rev3 [Jeff].sm (91kb) downloaded 24 time(s).

Genfit Hahn.sm (35kb) downloaded 22 time(s).



Offline djlischer  
#4 Posted : 22 March 2017 05:37:43(UTC)
djlischer

Rank: Newbie

Groups: Registered
Joined: 06/03/2017(UTC)
Posts: 2
United States
Location: Acton, MA

Hi Jean,
Thanks for the files. I like what you did with the initialization in the Rational Rev 3 file. I'll take a look at the other files too. Yes, I do know Mathcad. I use Mathcad 15 at work, but I don't have it at home. As I was picking up SMath, I thought about what functions I was missing and Given/Find and Minerr/Minimize were two that came to mind. That's how I stumbled on your Minimize samples.
Regards, Jeff
Users browsing this topic
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.