INTAR(1)模型及其推断研究

Inference for Integer-valued Threshold AR(1) Model

整值时间序列在时间序列分析里具有非常重要的地位,在各个领域都有其广泛地应用,它能够拟合实际生活中的离散时间序列.线性模型在生活中的各领域都有应用,然而在现实生活中,许多现象无法用线性时间序列去建模,因此提出了非线性模型.一类非常重要的非线性模型为门限自回归(TAR)模型,基于整值时间序列模型考虑门限自回归模型,由此提出了一个新的模型,即整值门限自回归(INTAR)模型,给出了该模型的基本性质,并利用最小二乘与极大似然方法对此模型中的参数进行了估计.

The integer time series models are widely used in many fields,it can fit discrete time series in actual life.The linear model has a good application in all fields.However,in real life,many phenomena cannot be used to simulate the linear time series,therefore,the nonlinear model is put forward,a kind of important nonlinear models is called the threshold autoregressive model,and the threshold autoregressive model is combined with the integer time series.This paper proposes a new model of the integer threshold autoregressive model.Some basic properties of this model is given.