Statistical Analysis of Poisson Conditionally Nonlinear Autoregressive Time Series by Frequencies-Based Estimators

被引:2
|
作者
Kharin, Yu [1 ]
Kislach, M. [1 ]
机构
[1] Belarusian State Univ, Res Inst Appl Problems Math & Informat, Minsk 220030, BELARUS
关键词
statistical forecasting; integer-valued time series; Poisson distribution; Markov chain; frequencies-based estimators; ROBUSTNESS; PARAMETERS; MODELS;
D O I
10.1134/S1054661820010083
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
Poisson conditionally nonlinear autoregressive model is proposed for integer-valued time series. Frequencies-based estimators (FBE) for model parameters, statistical forecasting statistics, and statistical tests for this model are constructed; their performance is analyzed theoretically and by computer experiments on simulated and real data.
引用
收藏
页码:22 / 26
页数:5
相关论文
共 50 条