Discrete-time ARMA;
Continuous-time ARMA;
CARMA;
Levy process;
Embedding;
D O I:
10.1016/j.jspi.2018.01.004
中图分类号:
O21 [概率论与数理统计];
C8 [统计学];
学科分类号:
020208 ;
070103 ;
0714 ;
摘要:
Given any stationary time series {X-n : n is an element of Z} satisfying an ARMA(p, q) model for arbitrary p and q with infinitely divisible innovations, we construct a continuous time stationary process {chi(t) : t is an element of R} such that the distribution of {chi(n) : n is an element of Z}, the process sampled at discrete time, coincides with the distribution of {X-n}. In particular the autocovariance function of {chi(t)} interpolates that of {X-n }. (C) 2018 Elsevier B.V. All rights reserved.