An estimate of the root mean square error incurred when approximating an f L^2( R) by a partial sum of its Hermite series

From MaRDI portal
(Redirected from Publication:1649150)
An estimate of the root mean square error incurred when approximating an \(f\in L^2(\mathbb R)\) by a partial sum of its Hermite series



Abstract: Let f be a band-limited function in L2(mathbbR). Fix T>0 and suppose fprime exists and is integrable on [−T,T]. This paper gives a concrete estimate of the error incurred when approximating f in the root mean square by a partial sum of its Hermite series. Specifically, we show, for K=2n,quadninZ+, left[frac{1}{2T}int_{-T}^T[f(t)-(S_Kf)(t)]^2dt ight]^{1/2}leq left(1+frac 1K ight)left(left[ frac{1}{2T}int_{|t|> T}f(t)^2dt ight]^{1/2} +left[frac{1}{2T} int_{|omega|>N}|hat f(omega)|^2domega ight]^{1/2} ight) +frac{1}{K}left[frac{1}{2T}int_{|t|leq T}f_N(t)^2dt ight]^{1/2} +frac{1}{pi}left(1+frac{1}{2K} ight)S_a(K,T), in which SKf is the K-th partial sum of the Hermite series of f,hatf is the Fourier transform of f, displaystyleN=fracsqrt2K+1+%sqrt2K+32 and fN=(hatfchi(−N,N))vee(t)=frac1piint−inftyinftyfracsin(N(t−s))t−sf(s)ds. An explicit upper bound is obtained for Sa(K,T).


Summary: Let \(f\) be a band-limited function in \(L^2(\mathbb R)\). Fix \(T>0\), and suppose \(f'\) exists and is integrable on \([-T,T]\). This paper gives a concrete estimate of the error incurred when approximating \(f\) in the root mean square by a partial sum of its Hermite series. Specifically, we show, that for \(K=2n\), \(n \in Z_+\), \(\left[ \frac{1}{2T},\int_{- T}^T,[f(t) -(S_K f)(t)]^2,d,t\right]^{1 / 2} \leq \left(1,+,\frac{1}{K}\right) \left(\left[\frac{1}{2 T},\int_{| t| > T},f,(t)^2,d,t\right]^{1/2},+,\left[\frac{1}{2 T},\int_{| \omega| > N},| \hat{f}(\omega)|^2,d,\omega\right]^{1/2}\right) + \frac{1}{K} \left[\frac{1}{2 T},\int_{| t| \leq T},f_N,(t)^2,d,t\right]^{1/2}+\frac{1}{\pi} \left(1,+,\frac{1}{2 K}\right) S_a(K, T),\) in which \(S_K f\) is the \(K\)-th partial sum of the Hermite series of \(f\), \(\hat{f}\) is the Fourier transform of \(f\), \(N=\frac{\sqrt{2K+1}+\sqrt{2K+3}}{2}\) and \(f_N = (\hat{f} \chi_{(- N, N)})^\vee(t) = \frac{1}{\pi} \int_{- \infty}^\infty \frac{\sin(N(t - s))}{t - s} f(s) d s\). An explicit upper bound is obtained for \(S_a(K, T)\).











This page was built for publication: An estimate of the root mean square error incurred when approximating an \(f\in L^2(\mathbb R)\) by a partial sum of its Hermite series

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1649150)