MATLAB Answers

Options when drawing using the Curve Fitting tool

2 views (last 30 days)
haeyeon JI
haeyeon JI on 22 Aug 2020
Commented: Alex Sha on 24 Aug 2020
I am inputting x y z data and using 'Curve Fitting tool' to create a nonlinear regression graph.
I am trying to find the coefficients a,b,c,d suitable for the data using the 'Custom Equation' formula, but the graph is too different from the data.
I tried changing Robust and Algorithm, but the graph is not similar to the data.
What additional work should I do to get the graph?


Sign in to comment.

Accepted Answer

Alex Sha
Alex Sha on 23 Aug 2020
Hi, post your data and function in txt format, if possible.


Show 1 older comment
Walter Roberson
Walter Roberson on 24 Aug 2020
The function in the image appears to be a*x*b*log10(y+c)+d ?
If so then a cannot be distinguished from b: you can only distinguish their product.
haeyeon JI
haeyeon JI on 24 Aug 2020
It is an expected expression,
and my expression is ax+blog10(y+c), not ax*blog10(y+c).
And please note that the data has been modified
Alex Sha
Alex Sha on 24 Aug 2020
The fitting result is not so good if the function is: z=a*x+b*log10(y+c)
Root of Mean Square Error (RMSE): 0.115584191368428
Sum of Squared Residual: 0.654625559420376
Correlation Coef. (R): 0.855538827407886
R-Square: 0.731946685202461
Adjusted R-Square: 0.714076464215958
Parameter Best Estimate
---------- -------------
a 0.0741121571659251
b -0.283418916981476
c -9.47213514559854

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!