Relative information. Theories and applications. With forewords by G. J. Klir and H. Haken (Q1188651)

This book is a comprehensive presentation of Jumarie's approach to information theory, which he developed during the last decade. His principal goal is to develop an information theory that incorporates not only syntactic aspects of information but also its semantic aspects. What is most interesting, however, is that he attempts to achieve this goal by an appropriate modification of the Shannon theory. One of the most significant contribution of this book is a careful comparison of discrete and continuous entropy. Jumarie's conclusion is, contrary to what is taken for granted by many scientists, that the continuous Shannon entropy is much more satisfactory than the discrete one. The main results of the book are: (1) a unified approach to discrete and continuous entropy, (2) a theory of subjective information based on a slight modification of Shannon's framework, (3) a theory of relative information derived from a model of observation with information invariance, (4) a uniform approach to various definitions of informational entropies which have been suggested in the literature, (5) some models of informational entropies for stochastic processes, (6) an introduction to a general theory of information without probability for deterministic functions, and (7) a basis for a generalization of the Bayesian statistical decision theory, which involves subjective parameters. The book is written in the spirit of physics, containing the various intuitions which led the physicists to the theory. In spite of the fact that some topics may be viewed as controversial, the book consists of highly original and thought provoking material and should be of interest to anyone interested in information theory and all connected research areas.