Random fitting when use fittype function in Toolboxes

Hi Matlab users

I am working on the data fitting by fittype function in Toolboxes. Basically, I have a set of data, x and y. I want to fit it with a function I defined. After I went through the document of fittype, I managed to get it work. However, I found that the fitting process is out of my control. I run the same code without changing anything, and the output of the fitted curve is quite different. The fitted results are listed below.Also, I got a warning from matlab "Start point not provided, choosing random start point." Not too sure they are related or not. My question is how to control the fitting process so that I can get the best fit like the first plot?

First run

Second run

Third run

Any help would be appriciated.

[Fan Li]

Thanks.

Answer 1

[Cris LaPierre]

They are absolutely related. It is the randomly selected start point that is leading to you getting different results each time. See this post.