| T O P I C R E V I E W |
| evanm |
Posted - 06/21/2004 : 06:21:03 AM
Hello,
I have been trying to perform a Lorentzian fit on a section of data but I have run into a few problems. My code looks as follows:
nlsf.begin();
nlsf.init();
nlsf.func$ = "Lorentz";
nlsf.fitdata1$ = %(%H,@D); //gets dataset of active graph
nlsf.dataBegin = Estarti+1; //starting index
nlsf.dataEnd = Eendi+1; //end index
nlsf.dataStep = 1;
nlsf.iterate(20);
//deal with fitting parameters here
nlsf.unInit();
nlsf.cleanUpFitData();
nlsf.end();
My script is designed to be ran and then the user clicks on the active graph from which I calculate the indices Estarti and Eendi. I have found that when I run this I get garbage values for the nlsf parameters (nlsf.p#). But then if I perform the same fit by clicking "Alanysis->Fit Lorentzian" and then rerun the script I get the propper values. Does anyone know why this is?
Thanks, Evan. |
| 3 L A T E S T R E P L I E S (Newest First) |
| evanm |
Posted - 06/22/2004 : 09:57:43 AM
Ok now after getting it to work seemingly just fine I have run into another problem. It basically goes as follows: I can run my fitting script sequentially as many times as I like and all is good. Then if I fit the same graph data using the menu it also works fine. But then if I run my script on the same graph I get the following error:
The number of cells in a matrix (rows X columns) must not exceed 90,000,000. Please re-enter your matrix row and column dimensions accordingly.
unknown error
unknown error
When this happens I also have the problem that a curve created from the graph's datasets is not created properly, so if I try to access its elements I get garbage values. ie...
Curve cvData(dsXName, dsYName);
Then if I do Curve_x(&cvData, 0) I get -0.0000 which makes no sense and is not the value in the worksheet. (And yes I have the dataset names correct). Does anyone know what is going on here, or has anyone ever had a problem like this before?
Thank you, Evan Morris. |
| evanm |
Posted - 06/22/2004 : 03:57:41 AM
Ok I figured out my problem. I simply had to provide the fitter with some initial fitting parameter values to help it along. This is why it worked after I fitted from the pull down menu. |
| Mike Buess |
Posted - 06/21/2004 : 8:24:38 PM
Hi Evan,
%C holds the name of the active dataset, so try this...
nlsf.fitdata1$=%C;
Mike Buess
Origin WebRing Member |
|
|
