loading page

The Gauss-cos model for the nonlinear approximation of autocorrelation functions
  • Zongmin Wu,
  • Ran Yang
Zongmin Wu
Fudan University
Author Profile
Ran Yang
Fudan University

Corresponding Author:[email protected]

Author Profile

Abstract

We derived a Gauss-cosh model for autocorrelation functions in the Fourier transform space by adopting the mass accumulation principle of the Grassmann space to the amplitude-frequency instead of the mass-point. Based on the Gauss-cosh model in the Fourier transform space, our main Gauss-cos model for autocorrelation functions was suggested on time domain. The physical interpretation of the Gauss-cos model was provided too. The new Gauss-cos model may be better for the kernel learning, pattern recognition or the approximation of the autocorrelation function to solve the time-dependent problems. Numerical experiments validated that the Gauss-cos model was efficient to fit the autocorrelation function of the fertility rate. The first two major principal phases of the Gauss-cos model we found for the autocorrelation function of the fertility rate were caused by the public diffusion and the mother-daughter diffusion.