Towards efficient fuzzy information processing. Using the principle of information diffusion (Q5960940)

The aim of the monograph is to present a comprehensive and self-contained theory of information diffusion as an important tool in fuzzy system analysis. The general idea is to develop suitable sound techniques to allow better understanding of the behavior of small samples by observing experiments. The developments are mainly focused on the use of the principle of information diffusion in partially filling the gaps caused by incomplete information. According to the mentioned inference paradigm, the recognition of the relationships between input and output can be improved by changing the crisp observations into fuzzy sets and consequently the solutions of the pattern recognition tasks could be significantly improved through the fuzzification of observations. The monograph contains a series of new concepts and techniques as for instance the concept of information matrix, the soft histogram estimate and an algorithm of the interior-outer-set model for fuzzy risk analysis. Also, a special attention is paid to risk issues such as the essence of risk, fuzzy risk definition and computation. The structure of the monograph is given by two parts containing twelve chapters. The first part supplies a theoretical treatment of the basic concepts: information distribution, information diffusion, quadratic and respectively normal diffusion. The second part is devoted to applications supporting the soundness of the considered theoretical attempts: estimation of epicentral intensity, estimation of isoseismal area, fuzzy risk analysis and calculation, analytic model for natural disasters. The plea on promoting the principle of information diffusion is convincing and the monograph can be recognized as an important step toward developing a new and promising branch of fuzzy set theory.