1
$\begingroup$

this is my first post so if I have any error while writing this, I'm sorry.

I had to do a polynomic interpolation for one of my lab experiments, and I need to get the error from it in order to continue. Any idea on how to do it?

This is the code:

f = {1.611 , 1.438 , 1.367 , 1.279 , 1.186 , 1.024 , 0.789 , 1.719 , 
   1.815 , 1.921 , 2.041 , 2.360 , 2.831};
i = {0.0453, 0.0412, 0.0383, 0.0342, 0.0288, 0.0216, 0.0142, 0.0433, 
   0.0400, 0.0367, 0.0321, 0.0237, 0.0173};
datos = Transpose[{f, i}]
interp = InterpolatingPolynomial[datos, x];
Expand[interp]

I supose every parameter from the interpolation will have it's own uncertainty, but what I don't know is if Mathematica is able to show them all to me. Thanks in advance

$\endgroup$
3
  • $\begingroup$ 5442.62 - 45353.6 x + 170957. x^2 - 385526. x^3 + 579400. x^4 - 611462. x^5 + 464716. x^6 - 256317. x^7 + 101841. x^8 - 28431.2 x^9 + 5294.38 x^10 - 590.532 x^11 + 29.8392 x^12 is wrong? This is my result for your code. $\endgroup$ Oct 25 '18 at 10:01
  • $\begingroup$ Are you looking for an exact interpolation (as you did) of degree 12 ? Or are you looking for a polynomial approximation of your data p? $\endgroup$ Oct 25 '18 at 10:04
  • $\begingroup$ "I suppose every parameter from the interpolation will have it's own uncertainty" -- There two sources, the round-off error in computing interp and the experimental error in f and i. I doubt a symbolic expression in terms of unknown individual errors in f and in i will be useful, but perhaps if the errors in f and i are each a known number or a known bound, then the uncertainty in the coefficients could be computed. $\endgroup$
    – Michael E2
    Oct 25 '18 at 13:40
1
$\begingroup$
 Total[(((interp /. x -> #)  & /@ f) - i)^2]

But the error is approximately zero. In abstract reality, this error is exactly 0, because you interpolate 13 values by the polynom of degree 12.

$\endgroup$

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.