On convergence of discrete methods of least squares on equidistant nodes

Abstract: We consider the well-known method of least squares on an equidistant grid with

N + 1

nodes on the interval

[- 1, 1]

with the goal to approximate a function

f i n m a t h c a l C l e f t [- 1, 1 i g h t]

by a polynomial of degree

n

. We investigate the following problem: For which ratio

N / n

and which functions do we have uniform convergence of the least square operator

{L S}_{n}^{N} : m a t h c a l C l e f t [- 1, 1 i g h t] i g h t a r r o w m a t h c a l P_{n}

? We investigate this problem with a discrete weighting of the Jacobi-type. Thereby we describe the least square operator

{L S}_{n}^{N}

by the expansion of a function by Hahn polynomials

. Without additional assumptions to functions

f i n m a t h c a l C l e f t [- 1, 1 i g h t]

it can not be guaranteed uniform convergence. But with

and additional assumptions to

f

and

l e f t (N_{n} i g h t)_{n i n m a t h b b N}

we obtain convergence and prove the following results: For an

a l p h a g e q 0

let

f i n l e f t g i n m a t h c a l C^{i} n f t y l e f t [- 1, 1 i g h t] : l i m l i m i t s_{n o i n f t y} s u p l i m i t s_{x i n [- 1, 1]} l e f t l v e r t g^{(n)} (x) i g h t v e r t f r a c n^{a l p h a + 1 / 2} 2^{n} n! = 0 i g h t

and let

(N_{n})_{n}

be a sequence of natural numbers with

N_{n} g e q 2 n (n + 1)

. Then the method of least squares

{L S}_{n}^{N_{n}} [f]

converges uniform on

[- 1, 1]

. Before we determine the maximum error ("worst case") with respect to the sup norm on the classes

m a t h c a l K_{n + 1} := l e f t f i n m a t h c a l C^{n + 1} l e f t [- 1, 1 i g h t] : s u p l i m i t s_{x i n [- 1, 1]} l e f t l v e r t f^{(n + 1)} (x) i g h t v e r t l e q 1 i g h t

.

This page was built for publication: On convergence of discrete methods of least squares on equidistant nodes