Centre Tecnològic de Nutrició i Salut (CTNS), TECNIO, CEICS, Reus, Spain.
Stat Med. 2011 Nov 20;30(26):3125-36. doi: 10.1002/sim.4336. Epub 2011 Aug 17.
We present a model based on two-order integer-valued autoregressive time series to analyze the number of hospital emergency service arrivals caused by diseases that present seasonal behavior. We also introduce a method to describe this seasonality, on the basis of Poisson innovations with monthly means. We show parameter estimation by maximum likelihood and model validation and show several methods for forecasting, on the basis of long-time means and short-time and long-time prediction regions. We analyze an application to model the number of hospital admissions per week caused by influenza.
我们提出了一个基于二阶整数自回归时间序列的模型，用于分析因呈现季节性行为的疾病而导致的医院急诊服务就诊人数。我们还介绍了一种基于每月均值泊松创新的方法来描述这种季节性。我们展示了最大似然法的参数估计和模型验证，并展示了几种基于长时间均值和短时间及长时间预测区域的预测方法。我们分析了一个应用案例，即用该模型来模拟每周因流感而住院的人数。
