Skip to main content
M
mzarrabi1-Visitor
1-Visitor
November 28, 2017
Question

Problem using minerr

  • November 28, 2017
  • 2 replies
  • 1694 views

Hi everybody,

I am trying to use minerr for optimization of my function which has 13 parameters. The problem is that although my function and my data are sinusoidal, the Weibull fit gives me a line instead of a sinusoidal curve passing through the points. I wonder if somebody could take a look at it. Many thanks in advance.

2 replies

R
RichardJ19-Tanzanite
19-Tanzanite
November 29, 2017

You have one equation and 13 unknowns. That is not solvable, even as a minimization problem.

L
LucMeekes23-Emerald IV
23-Emerald IV
November 29, 2017

I guess this is related to https://community.ptc.com/t5/PTC-Mathcad-Questions/Modification-and-or-calibration-function/m-p/492166#M174577

 

Your experimental results look much like an exponentially growing sinewave: f(N)=exp(N/No)*sin(N/Nt)

So it is understandable that it does not fit a Weibull function.

 Next to that, the D_M function behaves very ill (or am I abusing it?):

LM_20171129_ParamFit.png

This is not a type of function that you can do minerr or other types of parameter estimations with.

Luc

 

 

 

    M
    mzarrabi1-VisitorAuthor
    1-Visitor
    November 29, 2017

    Thank you for your responses @LucMeekes and @RichardJ. You are right Luc, my function was ill. I modified it and it works well now (see the attached). Exactly, that is an exponentially growing sinewave, but if you look at the pic, the weibull can estimate it somehow, but its amplitude is not high enough. Actually, it is related to the topic you mentioned. To tell you the truth, I did not understand what should I do finally.  data and function.JPGfunction.JPG

    Terms & ConditionsAccessibility statement