Springer Series in Statistics
1 total work
The area of estimating functions has wide applications and has undergone rapid development during the recent years. As a theory, it is sufficiently general to include most of the important aspects of the classical theory of statistical inference and to accomodate a variety of the more recent themes such as the Generalized Linear Models (GLM), the Generalized Estimating Equations (GEE), the Generalized Linear Mixed Effect Models (GLMM), the various forms of Autoregressive Conditionally Heteroscedastic models (ARCH), the Restrictive Maximum Likelihood (REML), the Empirical Likelihood, as well as many estsimators in Surivval Analysis, Nonparametric Regression, and Spatial Statistics. By the constructive nature of this theory, it will no doubt also provide us with the insights to derive new statistical procedures for scientific problems that will arise. Yet, these advantages do not make the subject more difficult to understand, for it is built on a handful of elementary tools such as linearization and projection, which apply repeatedly at different levels of sophistication.