In the Model Base Calibration toolbox, can I plot the coefficients from the surface response?
1 visualización (últimos 30 días)
Mostrar comentarios más antiguos
Sylvain
el 6 de Mayo de 2022
Respondida: Ian Noell
el 11 de Jul. de 2023
I am trying to figure out how to work with the MBC toolbox, especially to conduct Design of experiments.
I am testing a classic full fractional design, 2 levels (y = a0 + a1 x1 + a2 x2 + a3 x1.x2), taken from "Les Plans d'Experiences" Chapt 2, from Jacques Goupy.
I am quite succesful in plotting the response model: ![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/988940/image.png)
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/988940/image.png)
Where is the function to plot the coefficient value has an bar chart?
Also the workflow is very confusing, and need some help to understand how to better load my dataset: do i need to have coded/uncoded values? can I swap the input order?
Many thanks
0 comentarios
Respuesta aceptada
Ian Noell
el 11 de Jul. de 2023
MBC is intended as a tool for response surfaces for the purposes of prediction rather than analysing coefficients.
There is no bar chart for the coefficents.
The data should be in uncoded units. By default, the data range is used for coding.
The inputs are in alphabetical order when you use 'Fit Models'. If you create a new test plan using 'Design Experiment' instead of 'Fit Models' then you can define the inputs in any order you choose. You can also set the input ranges manually. There is a separate 'Fit Models' function in a test plan that allows you to fit models to data after designing an experiment.
0 comentarios
Más respuestas (0)
Ver también
Categorías
Más información sobre Model Assessment en Help Center y File Exchange.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!