群集成對資料之一致性的強韌推論;Robust likelihood inference for agreement between two procedures for clustered matched-pair data

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/74522

Title:	群集成對資料之一致性的強韌推論;Robust likelihood inference for agreement between two procedures for clustered matched-pair data
Authors:	梁琬琪;Liang, Wan-Chi
Contributors:	統計研究所
Keywords:	Kappa統計量;一致性;成對資料;Kappa statistic;agreement;matched-pair data
Date:	2017-06-22
Issue Date:	2017-10-27 14:00:56 (UTC+8)
Publisher:	國立中央大學
Abstract:	Kappa統計量是一種最常被用來測量評估者間一致性程度，或是不同篩檢方法間之一致性的統計量之ㄧ。最近，Yang and Zhou (2014) 提出了一種在群集成對資料下kappa的無母數變異數估計方法。本文致力於針對群集成對資料的kappa進行推論，建立kappa參數之強韌概似函數。儘管我們只在群集成對資料為兩個觀察值的情形下討論，此方法仍可適用於多個觀察值的群集資料中。利用此強韌概似函數方法，在不需特別針對群集間的相關性建構模型假設下，我們仍可得到概似比檢定統計量以及基於概似比檢定得到信賴區間等正確的推論工具。 ;Kappa statistic is one of the most utilized statistics for measuring agreement between raters or the consistency between screening devices for independent data. Recently, Yang and Zhou (2014) proposed a nonparametric variance estimate for kappa when data are collected in clusters. This work is devoted to the construction of robust likelihood inference for kappa for matched-pair data. In spite of the fact that we confine to the cases with clusters of size two, the methodology is applicable to general situations. We will derive tools including the likelihood ratio (LR) statistic and the LR test-based confidence interval that remain valid without specifically introducing or modeling the intra-cluster correlation.
Appears in Collections:	[Graduate Institute of Statistics] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	279	View/Open

社群 sharing

Loading...