Applications and methodology for multiplicative measurement error models and Bayesian model averaging in epidemiology /
Two general methodologies for epidemiologic studies are considered: multiplicative measurement error models and Bayesian model averaging. The methods axe examined in the context of problems in nutritional and genetical epidemiology, respectively. The nutritional application is based on the estimat...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1998.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=737708301&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | Two general methodologies for epidemiologic studies are considered: multiplicative measurement error models and Bayesian model averaging. The methods axe examined in the context of problems in nutritional and genetical epidemiology, respectively. The nutritional application is based on the estimation of a polynomial regression model in the presence of multiplicative measurement error in one of the predictors. Two general methods are considered, with the methods differing in their assumptions about the distributions of the predictor and of the measurement errors. Consistent parameter estimators and asymptotic standard errors are derived using estimating equation theory. Diagnostics axe presented for distinguishing between additive and multiplicative measurement error. Data from a nutrition study are analyzed using the methods. The results from a simulation study are presented and the performances of the methods compaxed. The genetical application is the estimation of a regression model that allows for correlation due to genetic relatedness. Bayesian model averaging is used to account for model uncertainty by using composites of posterior distributions to estimate parameters, with each posterior computed under a different model. The composites axe weighted averages, with weights taken to be model posterior probabilities. A computer simulation based on results from a prior study of vaxiation in the apolipoprotein B gene is used to compare this Bayesian approach to more conventional likelihood-based approaches. |
|---|---|
| Item Description: | Vita. "Major Subject: Statistics". |
| Physical Description: | xi, 64 leaves : illustrations ; 28 cm. Issued also on microfiche from University Microfilms Inc. |
| Bibliography: | Includes bibliographical references: pages 49-50. |