Seasonal count time series

Count time series are widely encountered in practice. As with continuous valued data, many count have seasonal properties. This paper uses a recent advance stationary to develop general modeling paradigm. The model constructed here permits any marginal distribution for the and most flexible autocorrelations possible, including those negative dependence. Likelihood methods of inference explored. first develops methods, which entail discrete transformation Gaussian process having dynamics. Properties this class then established particle filtering likelihood parameter estimation developed. A simulation study demonstrating efficacy is presented an application number rainy days successive weeks Seattle, Washington given.