Generalized endpoint-inflated binomial model

doi:10.1016/j.csda.2015.03.009

CSpace > 应用数学研究所

	Generalized endpoint-inflated binomial model
	Tian, Guo-Liang 1; Ma, Huijuan 2,4; Zhou, Yong3,4 ; Deng, Dianliang 5
	2015-09-01
发表期刊	COMPUTATIONAL STATISTICS & DATA ANALYSIS
ISSN	0167-9473
卷号	89 页码:97-114
摘要	To model binomial data with large frequencies of both zeros and right-endpoints, Deng and Zhang (in press) recently extended the zero-inflated binomial distribution to an endpoint-inflated binomial (EIB) distribution. Although they proposed the EIB mixed regression model, the major goal of Deng and Zhang (2015) is just to develop score tests for testing whether endpoint-inflation exists. However, the distributional properties of the EIB have not been explored, and other statistical inference methods for parameters of interest were not developed. In this paper, we first construct six different but equivalent stochastic representations for the EIB random variable and then extensively study the important distributional properties. Maximum likelihood estimates of parameters are obtained by both the Fisher scoring and expectation-maximization algorithms in the model without covariates. Bootstrap confidence intervals of parameters are also provided. Generalized and Fixed EIB regression models are proposed and the corresponding computational procedures are introduced. A real data set is analyzed and simulations are conducted to evaluate the performance of the proposed methods. All technical details are put in a supplemental document (see Appendix A). (C) 2015 Elsevier B.V. All rights reserved.
关键词	Endpoint-inflated binomial distribution Expectation-maximization algorithm Multinomial logistic regression model Stochastic representation Zero-inflated binomial distribution
DOI	10.1016/j.csda.2015.03.009
语种	英语
资助项目	National Natural Science Foundation of China (NSFC)[71271128] ; State Key Program of National Natural Science Foundation of China[71331006] ; NCMIS, Key Laboratory of RCSDS, CAS ; Shanghai University of Finance and Economics through Project 211 Phase IV, Shanghai First-class Discipline A ; Natural Sciences and Engineering Research Council of Canada ; IRTSHUFE, PCSIRT[IRT13077]
WOS研究方向	Computer Science ; Mathematics
WOS类目	Computer Science, Interdisciplinary Applications ; Statistics & Probability
WOS记录号	WOS:000357348000008
出版者	ELSEVIER SCIENCE BV
引用统计
文献类型	期刊论文
条目标识符	http://ir.amss.ac.cn/handle/2S8OKBNM/20255
专题	应用数学研究所
通讯作者	Ma, Huijuan
作者单位	1.Univ Hong Kong, Dept Stat & Actuarial Sci, Hong Kong, Hong Kong, Peoples R China 2.Univ Sci & Technol China, CM0FL, Hefei 230026, Anhui, Peoples R China 3.Chinese Acad Sci, Acad Math & Syst Sci, Beijing, Peoples R China 4.Shanghai Univ Finance & Econ, Sch Stat & Management, Shanghai, Peoples R China 5.Univ Regina, Dept Math & Stat, Regina, SK S4S 0A2, Canada
推荐引用方式 GB/T 7714	Tian, Guo-Liang,Ma, Huijuan,Zhou, Yong,et al. Generalized endpoint-inflated binomial model[J]. COMPUTATIONAL STATISTICS & DATA ANALYSIS,2015,89:97-114.
APA	Tian, Guo-Liang,Ma, Huijuan,Zhou, Yong,&Deng, Dianliang.(2015).Generalized endpoint-inflated binomial model.COMPUTATIONAL STATISTICS & DATA ANALYSIS,89,97-114.
MLA	Tian, Guo-Liang,et al."Generalized endpoint-inflated binomial model".COMPUTATIONAL STATISTICS & DATA ANALYSIS 89(2015):97-114.