Fourier inversion for multidimensional characteristic functions (Q1209215)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Fourier inversion for multidimensional characteristic functions |
scientific article |
Statements
Fourier inversion for multidimensional characteristic functions (English)
0 references
16 May 1993
0 references
Let \(f\) be a density function on \(\mathbb{R}^ n\), whose mean value function (i.e. \(r\in(0,+\infty)\to M_ rf(x)=\int_{S^{n- 1}}f(x+r\omega)d\omega\), where \(d\omega\) is normalized surface measure on the unit sphere) is piecewise \(C^ \infty\) and compactly supported. The characteristic function and its spherical partial sum are respectively defined by: \[ \hat f(\mu)=\int_{\mathbb{R}^ n}f(x)e^{i\langle\mu,x\rangle}dx,\qquad f_ R(x)={1\over (2\pi)^ n}\int_{|\mu|\leq R}\hat f(\mu)e^{- i\langle\mu,x\rangle}d\mu. \] Given a function \(f\) as above, some asymptotic results are proved; more precisely it is stated that \(\lim_{R\to+\infty}f_ R(x)=M_{0+}f(x)\) if and only if \(M_ rf\) has \(k=[(n-3)/2]\) continuous derivatives; otherwise \[ \liminf_{R\to+\infty}R^{-\nu}[f_ R(x)- M_{0+}f(x)]<0<\limsup_{R\to+\infty} R^{-\nu}[f_ R(x)- M_{0+}f(x)], \] where \(\nu\) \((\geq0)\) is uniquely determined. To achieve this result, some properties and asymptotic formulas for the spherical Bessel function and the Dirichlet kernel are proved. In conclusion some examples are discussed and some remarks concerning the Gibbs phenomenon are made.
0 references
characteristic function
0 references
spherical Bessel function
0 references
Dirichlet kernel
0 references
Gibbs phenomenon
0 references