الفهرس | Only 14 pages are availabe for public view |
Abstract Variable selection is one of the earliest topics that gain interest of statisticians. This interest is much less in the case of longitudinal data. Research on model selection for longitudinal data remains largely unexplored especially when the data is subject to measurement error and missingness that is the role not the exception for longitudinal data. Ignoring the existence of missing values in modelling, parameter estimation and variables selection in most of the cases lead to biased results. Covariates measurement error can negatively affect the accuracy of the estimates if not treated properly. Thus, in this thesis, we developed a simultaneous variable selection and parameter estimation method for longitudinal data that suffers from intermittent missing values and covariates measurement error. The penalized weighted generalized estimating equations and simulation selection extrapolation techniques are used. A simulation study is conducted to assess the method’s performance along with an application to LISS data. |