Avoid a cycle using lsqcurvefit (or lsqnonlin)

Hi everybody,

I have many curves of experimental data (n*1000) and I need to apply lsqcurvefit function to determine the coefficients of equations. As a result I expect two coefficients to each of my curve. So indeed I can use lsqcurvefit in a cycle to each of my curve. But if there is a way to apply this function without cycle?

Best regards

Answer 1

[Star Strider]

You would have to do your regressions iteratively in a loop.

I would ask lsqcurvefit for all the outputs, and then store them in a cell array (in a separate assignment line) in the event you find it necessary to calculate confidence intervals on the parameters or the fit, or otherwise compare the individual regressions from a statistical perspective.

[Star Strider]

Kerim Khemraev’s ‘Answer’ moved here —

Star Strider,

Does that mean that I cannot avoid a loop? Because that loop would take a lot of time

[Star Strider]

Yes.

Knowing only what you have described, you cannot avoid a loop.

[Star Strider]

Kerim Khemraev’s ‘Answer’ moved here —

Ok. Then is there another way to solve my problem with least-squares methods without a loop? What do you think?

[Star Strider]

It depends what your problem is.

I have no idea what it is, so I cannot provide a more exact solution.