具共變量誤差混合區間設限數據

機構典藏 > College of Sciences > Department of Applied Mathematics and Data Science > Research Paper > Item 987654321/102814

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/102814

Title:	具共變量誤差混合區間設限數據
Other Titles:	Mixed Case Interval-Censored Data with Covariates Errors
Authors:	溫啟仲
Contributors:	淡江大學數學學系
Keywords:	條件分數;區間設限;最大概似;測量誤差;自身一致;Conditional score;Interval censoring;Maximum likelihood;Measurement error;Self-consistency
Date:	2012-08
Issue Date:	2015-05-05 16:15:03 (UTC+8)
Abstract:	具共變量測量誤差之右設限數資料已被廣泛的研究，但對混合區間設限數據卻很少研究。研究動機為AIDS Clinical Trial Group (ACTG) 175 的研究，其中AIDS 的發生時間僅在斷續的回診時間才作觀察，而在進行治療前的共變量:基線CD4 數存有量測誤差。因此，我們對於具共變量測量誤差之混合區間設限資料，將提出一個結構模式的「半母數最大概似法」和一個機能模式的「條件分數法」來分析。當可以正確模式真實卻易有誤差之共變量的分布時，結構模式的方法是較有效的，而機能模式的方法無需模式真實共變量的分布，因此會較穩健。我們的目的為建立所提估計之漸進性質和計算法則。我們將藉由模擬和實例分析來評估和敘述統計方法。 Covariate measurement error problems have been extensively studied in the context of right-censored data but less so for mixed case interval-censored data. Motivated by the AIDS Clinical Trial Group (ACTG) 175 study, where the occurrence time of AIDS was examined only at intermittent clinic visits and the baseline covariate CD4 count was measured with error, we will describe a structural modeling method (semiparametric maximum likelihood approach) and a functional modeling method (conditional score approach) for analyzing mixed case interval-censored data with mismeasured covariates. The structural modeling method is more efficient when the distribution of the true covariate is specified correctly, while the functional modeling method is more robust due to it makes no distribution assumption on the true covariate. Our goal is to establish the asymptotic properties and computation methods of the proposed estimators. Besides, we will evaluate the methods through simulation studies and illustrate them with AIDS data.
Appears in Collections:	[Department of Applied Mathematics and Data Science] Research Paper

Files in This Item:

There are no files associated with this item.

Loading...