摘要
将STR类模型的转换函数设定为决定于某未知光滑有界函数的复合Logistic函数,提出半参数平滑转换回归模型.在独立同分布数据假设下,对其中的未知光滑有界函数采用级数估计,基于非线性最小二乘估计理论证明了参数估计量的相合性和渐近正态性,并简要讨论了置信区间的构造以及模型检验等问题.通过随机模拟与传统的STR模型进行比较,结果表明,该文的新模型及估计方法具有广泛的适用性和灵活性.
An unknown smooth function is substituted into the traditional smooth transition regression model and a semiparametric smooth transition regression model has been proposed in this paper.Based on the i.i.d.data assumption,we estimate the unknown smooth transition function by series estimator,the consistency and asymptotic normality properties of parameters are proved applying Nonlinear Least Square regression theory.The bootstrapping consistent confidence interval and hypothesis testing problem are also discussed briefly.The simulation results shows that,compared to traditional STR type model,our new model and estimating method are more flexible and have comprehensive applicability.
出处
《数学物理学报（A辑）》
CSCD
北大核心
2012年第4期797-807,共11页
Acta Mathematica Scientia
基金
教育部人文社会科学青年项目(10YJC790247)资助
关键词
平滑转换回归模型
级数估计
相合性
渐近正态性
随机模拟
Smooth Transition Regression Model； Series Estimator； Consistency； Asymptotic Normality； Simulation
