Characteristic function of a self-similar random process

Abstract

A stochastic differential equation is proposed for a characteristic function whose inverse function describes a self-similar random process with a power-law behavior of power spectra in a wide frequency range and a power-law amplitude distribution function. Gaussian "tails" for the characteristic distribution make it possible to evaluate its stability according to the formulas of classical statistics using the maximum of the Gibbs-Shannon entropy and, therefore, the stability of a random process given by an inverse function. Keywords: self-similar random processes, stochastic equations, power spectrum, 1/f-noise, maximum entropy.