Identifying Latent Components of the TINAR(1) Model

Abstract

In this paper we give a solution for the problem of identifying and predicting latent components of the integer-valued time series with skewed Skellam marginal distribution. At the beginning, expressions for latent components identification and prediction are derived. These expressions give us a possibility of revealing the values of two hidden, immeasurable components which affect the integer-valued time series with skewed Skellam marginals. Yule-Walker estimators of the unknown parameters are obtained. Also, quality of identification and lag-one prediction is tested on simulated data. At the very end, the model validation is performed in application on the real-life data.