In the mathematical subfield of numerical analysis, a trigonometric polynomial is a finite linear linear combination of sin(nx) and cos(nx) with n a natural number. Hence the term trigonometric polynomial as the sin(nx)s and cos(nx)s are used similar to the monomial basis for a polynomial.
The trigonometric polynomials are used in trigonometric interpolation to interpolate periodic functions. They are used in the discrete Fourier transform which is a special kind of trigonometric interpolation.
Let an be in C, 0 ≤ n ≤ N and aN ≠ 0 then
is called complex trigonometric polynomial of degree N. Using Euler's formula the polynomial can be rewritten as
Analogously let an, bn be in R, 0 ≤ n ≤ N and aN ≠ 0 or bN ≠ 0 then
is called real trigonometric polynomial of degree N.
Using the relation
- T2N(x) = eiNxtN(x)
we can construct a bijective mapping between the complex trigonometric polynomials and the real trigonometric polynomials. Thus a trigonometric polynomial can be considered a periodic function on the real line, with period some multiple of 2π, or as a function on the unit circle.
A trigonometric polynomial of degree N has a maximum of N roots in any open interval [a, a + 2π) with a in R.
A basic result is that the trigonometric polynomials are dense in the space of continuous functions on the unit circle, with the uniform norm. This is a special case, for example, of the Stone-Weierstrass theorem.