ABSTRACT

This paper brings forward a pth-order mixed dependence-driven random coefficient integer-valued autoregressive time series model (MDDRCINAR(p)). Stationarity and ergodicity properties of the proposed model are derived. The unknown parameters are estimated by conditional least squares, weighted least squares and maximum quasi-likelihood and asymptotic characterization of the obtained parameter estimators is proved. The performances of the proposed estimate methods are checked via simulations, which present that maximum quasi-likelihood estimators perform better than the other two estimate methods considering the proportion of within-Ω estimates in certain regions of the parameter space. The applicability of the model is investigated using two real count data sets.

摘要

本文提出了一种p阶混合依赖驱动的随机系数整数值自回归时间序列模型(MDDRCINAR( p ))。导出了所提出模型的平稳性和遍历性属性。通过条件最小二乘法、加权最小二乘法和最大拟似然法对未知参数进行估计，证明了所得参数估计量的渐近特征。所提出的估计方法的性能通过模拟进行了检查，这表明考虑到参数空间某些区域中Ω内估计的比例，最大似然估计比其他两种估计方法表现更好。使用两个真实计数数据集研究模型的适用性。