Spectral density estimation from random sampling for multiplicative stationary processes

In this paper, the spectral density estimation of a nonstationary class of stochastic processes is investigated. Although these processes are not stationary with respect to the additive binary operation, i.e., in the classical weak sense, they are stationary with respect to the multiplicative binary operation. These processes exist naturally as continuous-time processes. In order to answer many questions in practical situations using these processes, we develop a random sampling method for estimating their spectral densities by using a discrete-time process. Some simulation results are given.