Maximum Likelihood Estimation of ARMA Model with Error Processes for Replicated Observations

In this paper we analyse the repeated time series model where the fundamental component follows a ARMA process. In the model, the error variance as well as the number of repetition are allowed to change over time. It is shown that the model is identified. The maximum likelihood estimator is derived using the Kalman filter technique. The model considered in this paper can be considered as extension of the models considered by Anderson (1978), Azzalini (1981) and Wong and Miller (1990)