I have a simple function $x^n$ that I want to estimate in the interval of $[-1, 1]$ using polynomials of a lower degree. I understand this as having 3 possible kinds of best approximations:

  1. Least first power or L1 approximation.
  2. Least squares or L2 approximation.
  3. Chebyshev's or Minimax or $L_{\infty}$ approximation.

The goal is to make use of the 3rd kind of approximation to find the minimum max error that is obtained from some $$|f(x) - g(x)|$$

What I've done is:

  1. Normalise all the Chebyshev's polynomial so that the leading coefficient is 1. This means we let $$\bar{T}_k(x) = T_k(x) / 2^{k-1}$$

  2. Then we can find $x^n - \bar{T}_k(x)$ to have maximum degree of $n-2$ only.

But how do we use this property to economize a taylor series and estimate the error for one economization step?


Your Answer

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

Browse other questions tagged or ask your own question.