Nonparametric Methods in Change Point Problems

This volume deals with non-parametric methods of change point (disorder) detection in random processes and fields. A systematic account is given of up-to-date developments in this rapidly evolving branch of statistics. It also provides an approach to change point detection which is characterized by the reduction of change point problems to the more basic problem of mean value change points, and also the implementation of nonparametric statistics which require no a priori information concerning distributions. The book has seven chapters. The first presents an account of preliminary considerations; Chapter 2 reviews the current state-of-the-art. Chapters 3 and 4 - the major chapters of the book - consider a posteriori change point problems and sequential change point detection problems, respectively Chapter 5 discusses disorder detection of random fields, and Chapter 6 deals with applications in such diverse areas as geophysics, control systems and the analysis of historical texts. The volume concludes with a chapter devoted to new results, proofs and some technical details including an overview of a computer program package which has been developed for a posterior; change point detection. This volume is designed for researchers in the statistics and probability of random processes, and should also be of interest to specialists in control theory, engineering, systems analysis and cybernetics.