误差带有AR(q)的序列相关性的单指标模型估计

Estimation on the single index models with AR(q) serial correlated error

单指标模型由于具有高度的数据建模灵活性而在实际研究中应用广泛.基于B样条逼近技术和广义差分法,研究误差带有AR(q)的序列相关性的单指标模型最小二乘估计问题,并提出一套有效算法.通过大量数值模拟说明方法的有效性.

The single index models are widely used in many applications because of its flexible data modeling.In this paper,based on the Bspline approximation and generalized difference method,the least squares estimation was considered for the single index model with serial correlated random error.Through many simulation studies,it is illustrated that our proposed approach performs very well.