在模型錯誤下copula應用在迴歸分析時的正確性;The validity of copula for regression when model assumptions fail

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/77360

Title:	在模型錯誤下copula應用在迴歸分析時的正確性;The validity of copula for regression when model assumptions fail
Authors:	何建興;He, Jian-Xing
Contributors:	統計研究所
Keywords:	相關性資料;強韌概似函數;Copula;correlated data;robust likelihood function;copula
Date:	2018-06-29
Issue Date:	2018-08-31 14:35:19 (UTC+8)
Publisher:	國立中央大學
Abstract:	Copula 是目前分析相關性資料時，常用的一種建構聯合分配函數的一種方法，其優點在於可以便利地建構所需的聯合分配函數，但當Copula 模型假設錯誤時，其統計推論的正確性卻鮮有人探討。本文主要的目的是在廣義線性模型下，探討當Copula 模型假設錯誤時，其迴歸參數之估計量是否具有一致性，且能否提供正確的統計推論，並與Gamma-Gamma 模型、Poisson-Negative Binomial 模型以及Bivariate Negative Binomial 模型之結果做比對。;Copulas are popular and commonly used methods for constructing joint distribution functions when analyzing correlated data. The advantage of Copulas is that one can easily construct joint distribution functions with desired marginals. However, the validity of inference based on Copulas under model misspecification is rarely investigated. The objective of this paper is to examine the properties of the Copula-based estimates of the regression parameters given that the assumptions of the Copula model fail. We also make comparisons between several Copula models with other methods for analyzing bivariate correlated data.
Appears in Collections:	[Graduate Institute of Statistics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	207	View/Open

社群 sharing

Loading...