SMath Studio Forum
»
SMath Studio
»
Samples
»
Revising Levenberg Marquardt Method
Rank: Guest
Groups: Registered
Joined: 04/07/2015(UTC) Posts: 6,866 Was thanked: 981 time(s) in 809 post(s)
|
Carlos, my dear friend, SS 6179 has found his Waterloo from your latest LM algo style, but works fine from your original [2018] piece wise. Seemingly the matter is some numerical issue, but couldn't doctor that poisoned patient. You know: can't doctor a stone ... too dead for too long ! No more worry, all the fits in my collection are done. Still, LM is interesting. Cheers ... Jean, take care and always be good. Unsolved LM SS 6179.sm (135kb) downloaded 21 time(s).
|
|
|
|
Rank: Advanced Member Groups: Registered
Joined: 14/10/2015(UTC) Posts: 308
Was thanked: 77 time(s) in 58 post(s)
|
Dear friend Jean. The problem in the calculation of the second LM algorithm in SS6179 occurs in the calculation of the Jacobian matrix for finite differences, since in this step, after several iterations, for over unknown reasons an overflow error occurs (possibly a calculation engine error ). I have slightly modified the LMSA function and I have been able to solve the problem indicated with 4 data. The solution that uses all the data is not possible in SS6179. Solved LM SS 6179.sm (148kb) downloaded 25 time(s).I hope you feel comfortable with this. Best Regards Carlos
|
1 user thanked CBG for this useful post.
|
|
|
Rank: Guest
Groups: Registered
Joined: 04/07/2015(UTC) Posts: 6,866 Was thanked: 981 time(s) in 809 post(s)
|
Originally Posted by: CBG I hope you feel comfortable with this. Thanks Carlos for doctoring LMSA SS 6179. You are right: some numerical issues between versions. That was most noticed between the two SS 5346 Official and SS 5346 UNofficial ... this version cured all "Official" Included in the next Compendium Genfit Classroom Essentials Cheers ... Jean Genfit_11 Pharmacokinetics LMSA ... LM.sm (77kb) downloaded 18 time(s).
|
|
|
|
Rank: Guest
Groups: Registered
Joined: 04/07/2015(UTC) Posts: 6,866 Was thanked: 981 time(s) in 809 post(s)
|
|
|
|
|
Rank: Advanced Member Groups: Registered
Joined: 14/10/2015(UTC) Posts: 308
Was thanked: 77 time(s) in 58 post(s)
|
Originally Posted by: Jean Giraud Hola, Carlos. These two reconciliations [piecewise/global] work fine. Will test more on hard fit. Cheers ... Jean ... Thanks
Hola Jean. Thank you Excellent job, it was possible to put the rattle on the cat! but maybe it's time to migrate to the next version of SS. Best Regards Carlos
|
|
|
|
SMath Studio Forum
»
SMath Studio
»
Samples
»
Revising Levenberg Marquardt Method
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.