A Unified Nonparametric Test of Transformations on Cumulative Distribution Functions with Nuisance Parameters

(干扰参数存在下分布函数变换的统一非参检验)

 

主讲人:孙振庭北京大学国家发展研究院助理教授

主持老师:(北大国发院)黄卓

参与老师:(北大经院)王一鸣、王熙、刘蕴霆

(北大国发院)沈艳、张俊妮

(北大新结构)胡博

时间:20211210日(周五) 10:00-11:30

地点: 北京大学国家发展研究院承泽园246会议室

主讲人简介:

孙振庭,北京大学国家发展研究院助理教授。他于2018年从加州大学圣地亚哥分校获得经济学博士学位。他的研究方向是理论计量经济学和应用计量经济学。目前他的研究兴趣具体包括微观计量经济学,非参和半参计量模型,以及高维大数据模型。

摘要:

This paper provides a simple unified approach of testing transformations on cumulative distribution functions with nuisance parameters. We transform the hypothesis of interest into an equality restriction of a map on the differences of cumulative distribution functions. By showing that this map satisfies particular differentiability conditions, we establish the asymptotic distribution of the test statistic under null. We approximate this asymptotic distribution using a numerical bootstrap approach and construct the critical value for the test. The proposed test is shown to be asymptotically size controlled and consistent. Monte Carlo simulations show that the test performs well on finite samples.

Baidu
sogou