Time series models with tailed generalized geometric Linnik distribution as Marginals

Abstract

Tailed distributions are found to be useful in the study of life testing experiments and clinical trials. Tailed forms of type I and type II generalized geometric Linnik distribution and their asymmetric forms are studied in [1]. The usual technique of transferring data to use a Gaussian model fails in certain situations. Hence a number of non-Gaussian autoregressive models have been introduced by various researchers. A first order autoregressive model with tailed type I generalized geometric Linnik distribution is introduced in this paper. It is shown that the process is not time reversible. The model is extended to higher order cases.