Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let's have the discrete periodic function $f(t)$ which has only 5 non-zero values $f(t_0) = -10, f(t_1) = -5, f(t_2) = -10, f(t_3) = -15$ and $f(t_4) = -10$, all other points within the period $[t_0,t_4]$ being zero. The obvious interpolating function which recovers these values is $f(t) = 5sin(t) - 10$. Is this a unique interpolating function or you can propose other interpolating functions, such as polynomial functions with constraints? Any concrete examples, please?

share|cite|improve this question

Your $f$ only interpolates these values if $t_i=\frac{i\pi}{2}$. Also, $f$ is clearly not zero at all the other points in $[t_0,t_4]$.

Moreover note that you can, for any finite number of value pairs $(t_i,v_i)_i$, find arbitrarily many smooth functions $F$ s.t. $F(t_i)=v_i$ for all $i$. Look for example at Lagrange polynomials/Newton polynomials etc. You can always change the function values in between the prescribed points so that your function is still continuous or even smooth.

When you drop the continuity/smoothness requirement (which you do if you say that it should have only 5 non-zero values) then finding an interpolation function is completely trivial (just take the function that is 0 everywhere except on the prescribed points where it takes the prescribed values). If you necessarily want to write it down as a formula you can always use a linear combination of Kronecker deltas:

$$f(t)=\sum_i y_i \delta^t_{t_i}.$$

where $(t_i,y_i)_i$ is your data and $\delta^t_s$ is 1 if $t=s$ and 0 if not.

share|cite|improve this answer
So, then, what is $f'(t)$ of that function? Will it differ from $f'(t)$ when it is defined as $f(t) = 5sin(t_i) - 5$ where $i = 0, 1, 2, 3, 4$? – ganzewoort Oct 20 '12 at 14:03
It doesn't make sense to write down the derivative of a function that is only nonzero at finitely many points. It's not differentiable. However, its differentiable almost everywhere and the derivative is 0 almost everywhere. – Your Ad Here Oct 20 '12 at 14:04
This is the main point of my asking. If the interpolating function indeed is $f(t) = 5sin(t) - 10$ then it does have first derivative which is $5cos(t)$ and can be evaluated at the said 5 points. Also, from your answer I take it that there is no polynomial which would recover exactly the 5 points. All one can achieve through a polynomial is an approximation, right? – ganzewoort Oct 20 '12 at 14:09
Then you should have mentioned that in your question. Your function $f$ does NOT have only 5 non-zero values by the way. I think you got some things confused. – Your Ad Here Oct 20 '12 at 14:12
Of course there is a polynomial which recovers exactly these points. How did I say otherwise? Please look up Lagrange polynomials. Interpolating functions are not unique in the sense you seem to have in mind. – Your Ad Here Oct 20 '12 at 14:13

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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