函数型Probit模型的sieve极大似然估计  

Sieve Maximum Likelihood Estimator for Functional Probit Model

作　　者：王龙兵 杜江[1] 张忠占[1] WANG Long-bing;DU Jiang;ZHANG Zhong-zhan(College of Applied Sciences,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]北京工业大学应用数理学院

出　　处：《数学的实践与认识》2018年第18期229-235,共7页Mathematics in Practice and Theory

基　　金：国家自然科学基金(11771032)

摘　　要：研究函数型Probit模型的sieve极大似然估计的渐近性质．在一定的条件下，证明了估计的强相合性和渐近正态性以及该估计的非参数部分达到最优收敛速度．最后给出了一个模拟研究，表明sieve极大似然估计有较好的有限样本性质．This article studies the asymptotic properties of sieve maximum likehood esti-mator for functional Probit model. Under some mild conditions, it is proved that the stove MLEs are strong consistent and asymptotically normally distributed, and the nonparametric part of the sieve MLEs has an optimal convergence rate. Finally, a simulation study is carried out, which shows that the sieve MLEs have good finite sample properties.

关 键 词：函数型数据 PROBIT模型 sieve极大似然估计 渐近性质 强相合 最优收敛速度 

分 类 号：O212.1[理学—概率论与数理统计]