SOLVE BLOCK

[alexrue-disable]

A USUALLY HAVE THOUSANDS OF OBSERVATIONAL OR EXPERIMENTAL POINTS WHICH I WOULD LIKE TO FIT TO AN EQUATION TO FIND THE BEST VALUES FOR THE COEFFICIENTS.

THIS I CAN DO WITH A SOLVE BLOCK, AS LONG AS I DO NOT HAVE MANY POINTS, BECAUSE IT APPEARS THAT I MUST WRITE A DIFFERENT EQUATION FOR EACH POINT.

IS THERE A WAY TO USE A SOLVE BLOCK FOR MANY MANY (PERHAPS THOUSANDS) POINTS WITHOUT HAVING TO WRITE MANY MANY EQUATIONS???

HAS SOMEONE DONE THIS OR KNOW HOW TO DO IT?????????

CHEERS


ALEX

[TomGutman]

Don't type in all caps. It is hard to read, and the equivalent of shouting. Considered rather rude.

Minerr works fine with vector equations. Search th collaboratory for examples of curve fitting using minerr. There are lots of them.
__________________
� � � � Tom Gutman

[RichardJ]

Unfortunately, although what Tom says is true, there are so many examples of using minerr in the Collaboratory that finding a good, easy to understand one, is not necessarily easy. You can find a simple example if you go to "help", quicksheets", "Data Anlaysis", "Using minerr for nonlinear least squares fitting".

If you can't get it to work after reading that, post a worksheet containing the data and the function you want to fit.

Richard

[JCBanks]

Consider using a LS-Fit program such as
TableCurve 2D and TableCurve 3D.

Sincerely,

Julio C. Banks, MSME, P.E.

[AlvaroDíaz]

Hope attached helps.

Saludos a la UNAM. Alvaro.

[Fred_Kohlhepp]

Mathcad has a number of curve fitting routines built in. Rather than a solve block, see if one of those will perform. You can also use GENFIT if you know the form of the fitting curve and it's not standard.

Don't forget; Given/Find will report an answer ONLY if it cah get inside the TOL level for errors. If you're trying for a least-squares fit, a better choice would be Given/Minerr.


Fred Kohlhepp
fkohlhepp@sikorsky.com

[TomGutman]

There is little point to genfit. Minerr is easier to set up, more flexible, and just as fast and accurate.
__________________
� � � � Tom Gutman

[ptc-1368288]

Your demand could not be honored as you had no data set to fit. These two attached do sum a good % of the fitting techniques currently used by the "Mathcad Power Users", over 120 pages of modules to copy/paste. Another ! huge ! collection is pending " Mathcad for Engineers_Models", maybe next year sometimes. For each of the attached, follow strictly the directives at the top of the sheet.



jmG

[ptc-1368288]

... this is a good example that went in the "Lost Worlds",
Maybe like yours, if you don't revisit, attach the data set, scope.
3 collaborations = no reply !
+ & +, major collaborators wonder if they help in the blue ?

jmG