Chaos: A statistical perspective (Q5943422)

The focus of the book is on the interface between statistics and chaos. Thus, an informal approach is adopted. Four case studies are presented in the first chapter to illustrate the practical scope of the techniques and methods resulting from the statistical perspective to chaos. Chapter 2 explains some theoretical aspects of deterministic systems. The main concepts of deterministic chaos are intoduced: Lyapunov exponent, correlation dimension, delay coordinates and dynamics reconstruction. Then, stochastic systems are dealt with in chapter 3. Here the points addressed are: Ergodicity, shadowing, initial-value sensitivity, dimension of state vector, and observational noise. The next two chapters are on statistical analysis. They start with plots useful to detect chaotic behaviour. Tests for nonlinearity and estimation of the embedding dimension are other main topics of chapter 4. The sections of chapter 5 are: Correlation Dimension, Estimation of Measures of Initial-Value Sensitivity, and Map Reconstruction. Linear prediction fails to recognise that how well one can predict depends on where one is in the parameter space. Therefore chapter 6 deals with nonlinear least-squares prediction. After some theory the authors give several examples. Then they present a selective summary of recent approaches. Miscellaneous topics are touched next. They are of varying degrees of importance in the study of deterministic chaos and were selected to give a general flavour of the potential points of contacts with statistics. Finally, the appendix gives a more systematic introduction to some of the fundamentals of the theory of deterministic chaos, which complements the informal approach given in chapter 2.