1
$\begingroup$

I am curious about MATLAB function. Can you tell me why does not it do good approach to using of MATLAB function except for speed reason?

Is there any other reason to using polynomial utility?

$\endgroup$
  • $\begingroup$ what MATLAB function are you referring to that does polynomial interpolation? resample()? $\endgroup$ – robert bristow-johnson Aug 10 '15 at 0:44
2
$\begingroup$

It is basically an approach choice.
Inside the math is identical.

Usually, when doing Least Squares curve fitting, you're not looking for the Polynomial coefficients but a scaled version of them.
For instance, if you try to estimate Range, Velocity and Acceleration from Range Measurements using LS you need to scale the Polynomial Coefficients according to the dynamic model.

By the way, most of the times, you can do the LS estimation using $ {\left( {H}^{T} H \right)}^{-1} {H}^{T} $ which would be faster than pinv.
The only case it won't work is when $ \left( {H}^{T} H \right) $ isn't invertible (Rarely occurs in real world measurements).

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks Sir, Did you mean that we use pinv only in this ? The only case it won't work is when (HTH) isn't invertible (Rarely occurs in real world measurements). $\endgroup$ – gmotree Aug 9 '15 at 6:19
  • $\begingroup$ Yea, pinv uses the SVD which means it can generate the Pseudo Inverse of any matrix. Please mark the question as answered. $\endgroup$ – Royi Aug 9 '15 at 7:02
  • $\begingroup$ Thanks i have marked. BTW can you tell me which part are needed the curve fitting tech? $\endgroup$ – gmotree Aug 9 '15 at 7:06
  • $\begingroup$ I usually use $ {\left( {H}^{T} H \right)}^{-1} {H}^{T} $. $\endgroup$ – Royi Aug 9 '15 at 7:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.