State space modeling of long-memory processes

摘要: This paper develops a state space modeling for long-range dependent data. Although process has an infinite-dimensional representation, it is shown that by using the Kalman filter, exact likelihood function can be computed recursively in finite number of steps. Furthermore, approximation to based on truncated equation considered. Asymptotic properties these approximate maximum estimates are established class models, namely, fractional autoregressive moving average models. Simulation studies show rapid converging approach.