中科院国家数学与交叉科学中心
数学与物理交叉部
学术报告会
题 目：Superresolution methods of spectral analysis
报告人： Prof. Vladimir A. Mandelshtam（Univ. of California, Irvine）
时 间：07.08 (星期三) 10:00--11:00
地 点：数学院南楼N820室
摘 要:
Harmonic inversion is a central problem of spectral analysis of time signals, which arises, particularly, in various areas of Physics and Chemistry. It corresponds to a parametric fit of a time signal by a sum of complex sinusoids. I will introduce the Filter Diagonalization Method(FDM), which is a numerically efficient way to solve the harmonic inversion problem by mapping it to a generalized eigenvalue problem. A multidimensional version of the harmonic inversion problem also exists, and, consequently, the multidimensional FDM has been developed and applied successfully to data processing in Nuclear Magnetic Resonance experiments. Other related problems, such as multi-exponential fit, and related methods, will also be discussed.
