Unified methods for censored longitudinal data and causality (Q1858123)

This book provides a general estimating function methodology for locally semi-parametric efficient estimation of smooth parameters of large non- or semi-parametric models for the law \(F_X\) of very high dimensional data \(X\) based on the observed data \(Y\), when the data are coarsened at random. The text is divided into six chapters. Chapter 1 contains an overview of the mathematical tools of the subject, including the modern theory of semiparametric models based on tangent spaces and Hilbert space projections on these spaces. Chapter 2 provides a complete treatment of the locally efficient estimating function methodology and its associated asymptotic theory. Chapters 3 through 5 treat various censored data: right-censored, current status data, a combination of current status data and right-censored data and also multivariate right-censored data. Chapter 6 provides a general methodology for estimation of casual and non-casual parameters from longitudinal data, complicated by right censoring of some responses and interval censoring of other responses. Examples of the data analyses and simulations are provided at the end of each chapter.