The theory and applications of statistical inference functions (Q1210780)

This monograph gives a self-contained development of the theory of estimating functions with applications to estimation, censoring, robustness and inferential separation of parameters. In the introductory chapter of the book the authors explain the approach to the problem of estimating an unknown function which possesses the true value of an unknown parameter rather than estimating the parameter as a root, directly with reference to the maximum likelihood method of estimation. The material is covered in five chapters. The concepts of E-ancillarity and E-sufficiency, which permit inferential reduction analogous to the usual sufficiency and ancillarity reductions, are introduced in Chapter 2. Criteria for the selection of an inferential function with regularity conditions are given in Chapter 3. It is stressed here that the regularity is required of the function rather than the model. The concept of E-sufficiency in various settings with censoring problems and nuisance parameters is discussed in Chapters 4 and 5. The general theory developed in Chapter 2 is applied to problems involving parameters of stochastic processes in Chapter 6. The monograph underlines the idea of imposing regularity conditions on the class of allowable procedures, and then to search for an appropriate inference within the class.