The purpose of this book is to presenta methodology for designing and tuning fuzzyexpert systems in order to identify nonlinear objects; that is, to build input-output models using expert and experimental information. The results of these identifications are used for direct and inverse fuzzy evidence in forecasting and diagnosis problem solving.
The book is organised as follows: Chapter 1 presents the basic knowledge about fuzzy sets, genetic algorithms and neural nets necessary for a clear understanding of the rest of this book. Chapter 2analyzesdirect fuzzy inference based on fuzzy if-then rules. Chapter 3is devoted to the tuning of fuzzy rulesfor direct inference using genetic algorithms and neural nets. Chapter4presents models and algorithms for extracting fuzzy rules from experimental data. Chapter 5 describesa method for solving fuzzy logic equationsnecessary for the inverse fuzzy inference indiagnostic systems. Chapters6 and 7 aredevoted to inverse fuzzy inferencebased on fuzzy relations andfuzzy rules. Chapter 8presents a method for extracting fuzzy relations from data. Allthe algorithms presented in Chapters 2-8 arevalidated by computer experiments and illustrated bysolving medical and technicalforecasting anddiagnosis problems. Finally, Chapter 9includes applications of the proposed methodology in dynamicand inventory control systems, prediction of results of football games,decisionmaking in road accident investigations, project management and reliability analysis.
Table of Contents
Preface.- Fundamentals of intellectual technologies.- Direct inference based on fuzzy rules.- Fuzzy rules tuning for direct inference.- Fuzzy rules extraction from experimental data.- Inverse inference based on fuzzy relational equations.- Inverse inference with fuzzy relations tuning.- Inverse inference based on fuzzy rules.- Fuzzy relations extraction from experimental data.- Applied fuzzy systems.