東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Measurement error in covariates in t...

Greene, Wendy Faye.

FindBook

Google Book

Amazon

博客來

Measurement error in covariates in the marginal failure time model for multivariate data.

紀錄類型:	書目-語言資料,印刷品 : Monograph/item
正題名/作者:	Measurement error in covariates in the marginal failure time model for multivariate data./
作者:	Greene, Wendy Faye.
面頁冊數:	194 p.
附註:	Director: Jianwen Cai.
Contained By:	Dissertation Abstracts International61-11B.
標題:	Biology, Biostatistics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9993313
ISBN:	0493015760

Measurement error in covariates in the marginal failure time model for multivariate data.
Greene, Wendy Faye.

Measurement error in covariates in the marginal failure time model for multivariate data. - 194 p.

Director: Jianwen Cai.

Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2000.

When applying statistical models, statisticians frequently assume that the covariates of interest are known without error. For many reasons such as logistical infeasibility, failings of human memory, underdeveloped technology and financial limitations, statistical analyses must use imprecisely measured covariates, violating a modeling assumption. Measurement error introduces noise into the modeling process and biases the parameter estimates for the covariate measured with error and for other covariates in the model.

ISBN: 0493015760Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Measurement error in covariates in the marginal failure time model for multivariate data.
LDR:03768nam 2200325 a 45 001 929308
005 20110427
008 110427s2000 eng d
020 $a 0493015760
035 $a (UnM)AAI9993313
035 $a AAI9993313
040 $a UnM $c UnM
100 1 $a Greene, Wendy Faye. $3 1252792
245 1 0 $a Measurement error in covariates in the marginal failure time model for multivariate data.
300 $a 194 p.
500 $a Director: Jianwen Cai.
500 $a Source: Dissertation Abstracts International, Volume: 61-11, Section: B, page: 5657.
502 $a Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2000.
520 $a When applying statistical models, statisticians frequently assume that the covariates of interest are known without error. For many reasons such as logistical infeasibility, failings of human memory, underdeveloped technology and financial limitations, statistical analyses must use imprecisely measured covariates, violating a modeling assumption. Measurement error introduces noise into the modeling process and biases the parameter estimates for the covariate measured with error and for other covariates in the model.
520 $a Prentice (1982) first addressed measurement error in the univariate Cox model. He examined the causes of the bias and subsequent researchers explored the size of the bias and methods for accounting for measurement error (Pepe, Self and Prentice, 1989; Hughes, 1993; Zhou and Pepe, 1995; Wang, Hsu, Feng, and Prentice, 1997; Hu, Tsiatis and Davidian, 1998). These univariate methods assume that each failure time is taken from independent subjects. When the assumption of independence among subjects fails, survival analysis methods must account for the dependence among failure times. Marginal models have been suggested for multivariate failure time data when the correlation among the failure times is not of interest.
520 $a Although some methods have been proposed for dealing with measurement error in univariate failure time models, research that corrects for measurement error in multivariate failure time models has been limited. In this dissertation we determine the bias that results when the analyst naively substitutes the covariate measured with error into a standard marginal failure time model for multivariate survival data. We provide an expression which relates the true parameter estimate with the naive estimate.
520 $a We propose to use the SIMulation EXtrapolation (SIMEX) (Carroll, Ruppert and Stefanski, 1995) procedure to account for measurement error in covariates in marginal failure time models for multivariate failure time data. The advantage of the SIMEX approach is that it requires no assumptions about the distribution of the true underlying covariate. The method first SIMulates increasing amounts of measurement error and then EXtrapolates back to the case of no measurement error. We show that the SIMEX estimator is consistent and asymptotically normally distributed with a covariance matrix which may be consistently estimated. We also determine the asymptotic normal distribution of the estimator when the variance of the measurement error is not known but must be estimated. Our simulation studies demonstrate that the SIMEX method performs well in practical sample sizes. We also establish the validity of the SIMEX method for the univariate Cox model. We apply the proposed methods to data from the Hemophilia Growth and Development Study.
590 $a School code: 0153.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Health Sciences, Public Health. $3 1017659
650 4 $a Statistics. $3 517247
690 $a 0308
690 $a 0463
690 $a 0573
710 2 0 $a The University of North Carolina at Chapel Hill. $3 1017449
773 0 $t Dissertation Abstracts International $g 61-11B.
790 $a 0153
790 1 0 $a Cai, Jianwen, $e advisor
791 $a Ph.D.
792 $a 2000
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9993313