Structure of continuous-time ARMA process driven by semi-Levy measure

Abstract

A class of continuous-time autoregressive moving average (CARMA) process driven by simple semi-Levy measure is defined and its properties are studied. We discuss some new insights on the structure of the semi-Levy measure which is described as periodically divisible measure. This consideration enable us to provide statistical property of the introduced process. We show that this process is well defined without having to assume further conditions on the measure. We find a kernel representation of the process and present the properties of first and second moments of it. Finally we show the efficiency of our model by implying simulated data.