Integer-valued time series and renewal processes

摘要: This research proposes a new but simple model for stationary time series of integer counts. Previous work in the area has focused on mixture and thinning methods links to classical autoregressive moving-average difference equations; contrast, our use renewal process generate correlated sequence Bernoulli trials. By superpositioning independent copies such processes, with binomial, Poisson, geometric, or any other discrete marginal distribution can be readily constructed. The class proposed is parsimonious, non-Markov, generates either short long memory autocovariances. fitted linear prediction techniques series. Estimation parameters based conditional least squares considered. Asymptotic properties estimators are derived. models sometimes have an structure we consider AR(1) count case detail. Unlike previous tactics, negative autocorrelations produced.