header
Image from OpenLibrary

Complete case analysis and multiple imputations for longitudinal data with missing response and covariates / Nesma Mady Mohamed Darwish ; Supervised Ramadan Hamed Mohamed , Ahmed Mahmoud Gad

By: Contributor(s): Material type: TextTextLanguage: English Publication details: Cairo : Nesma Mady Mohamed Darwish , 2020Description: 124 P . : charts ; 25cmOther title:
  • تحليل البيانات الكاملة والتعويض المتعدد للبيانات الطولية فى حالة فقد المتغير التابع والمستقل [Added title page title]
Subject(s): Available additional physical forms:
  • Issued also as CD
Dissertation note: Thesis (Ph.D.) - Cairo University - Faculty of Economics and Political Science - Department of Statistics Summary: Longitudinal data with dropout are common in practice. Missing data indicate that the intended measurements for an individual are not available. There are two patterns of missing data: monotone and non-monotone missingness. In the same time, there are three types of missing data mechanisms: missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR). The difference between these types is related to the assumptions about whether the missingness depends on the observed and/or the unobserved responses. Incomplete longitudinal data can be modeled using complete case analysis which is a method that removes observations with any missing value from the analysis. Multiple imputation (MI) methods refer to replacing the missing values with a set of M plausible values. Replacing the missing values by a set of imputed values creates a complete data set. Replacing the missing values by another set of the imputed values creates another complete data set. Each imputed data set can be analyzed using the classical methods that assume the data set is complete.Biomedical research is plagued with problems of missing data, especially in clinical trails of medical and behavioral therapies adopting longitudinal design. After comprehensive literature review on modeling incomplete longitudinal data based on the full {u2013} likelihood functions, this dissertation proposes multiple imputation methods to deal with monotone missingness in cross-sectional covariates and in longitudinal response with dropout and modeling missing in the longitudinal response through a shared parameter model. There is a comparative study between the complete case analysis (CCA) , the proposed multiple imputation method to missing at random in longitudinal response and monotone missing in cross-sectional covariates and compare with non-random missingness in longitudinal response and deal with it through shared parameter model with multiply imputed cross-sectional covariates. This can be done through simulation studies and application study
Tags from this library: No tags from this library for this title. Log in to add tags.
Star ratings
    Average rating: 0.0 (0 votes)
Holdings
Item type Current library Home library Call number Copy number Status Date due Barcode
Thesis Thesis قاعة الرسائل الجامعية - الدور الاول المكتبة المركزبة الجديدة - جامعة القاهرة Cai01.03.01.Ph.D.2020.Ne.C (Browse shelf(Opens below)) Not for loan 01010110082485000
CD - Rom CD - Rom مخـــزن الرســائل الجـــامعية - البدروم المكتبة المركزبة الجديدة - جامعة القاهرة Cai01.03.01.Ph.D.2020.Ne.C (Browse shelf(Opens below)) 82485.CD Not for loan 01020110082485000

Thesis (Ph.D.) - Cairo University - Faculty of Economics and Political Science - Department of Statistics

Longitudinal data with dropout are common in practice. Missing data indicate that the intended measurements for an individual are not available. There are two patterns of missing data: monotone and non-monotone missingness. In the same time, there are three types of missing data mechanisms: missing completely at random (MCAR), missing at random (MAR), and missing not at random (MNAR). The difference between these types is related to the assumptions about whether the missingness depends on the observed and/or the unobserved responses. Incomplete longitudinal data can be modeled using complete case analysis which is a method that removes observations with any missing value from the analysis. Multiple imputation (MI) methods refer to replacing the missing values with a set of M plausible values. Replacing the missing values by a set of imputed values creates a complete data set. Replacing the missing values by another set of the imputed values creates another complete data set. Each imputed data set can be analyzed using the classical methods that assume the data set is complete.Biomedical research is plagued with problems of missing data, especially in clinical trails of medical and behavioral therapies adopting longitudinal design. After comprehensive literature review on modeling incomplete longitudinal data based on the full {u2013} likelihood functions, this dissertation proposes multiple imputation methods to deal with monotone missingness in cross-sectional covariates and in longitudinal response with dropout and modeling missing in the longitudinal response through a shared parameter model. There is a comparative study between the complete case analysis (CCA) , the proposed multiple imputation method to missing at random in longitudinal response and monotone missing in cross-sectional covariates and compare with non-random missingness in longitudinal response and deal with it through shared parameter model with multiply imputed cross-sectional covariates. This can be done through simulation studies and application study

Issued also as CD

There are no comments on this title.

to post a comment.