Problem with Curve Fitter 95% CI Upper and Lower Bounds' Equations

Attached is a screenshot that shows the coefficients (a, b, and c) of the upper and lower 95% CI bounds are not calculated correctly.

However, the graphical representation (after activating the prediction bounds) displays a correct 95% CI. The correct equations of the 95% CI are added at the bottom in red.

For example, for the upper bound, the y-intercept is reported to be 3.55, but we can see on the graph that it is 5.28.

Does anyone know what might be the reason for that?

Thank you.

Answer 1

[Torsten]

Have a read here:

https://uk.mathworks.com/help/curvefit/confidence-and-prediction-bounds.html

It will show you the difference between confidence bounds on coefficients and prediction bounds on fits.

I guess you mean the nonsimultaneous functional bounds:

p21 = predint(fitresult,x,0.95,'functional','off')
plot(fitresult,x,y), hold on, plot(x,p21,'m--')