Open AccessMORPHOGENESIS OF RULEBASES OF FUZZY INFERENCE SYSTEMS IN TERMS OF CONTRADICTOR EXPERT OPINIONS
Author(s) -
А.А. Сорокин
Publication year - 2019
Publication title -
vestnik astrahanskogo gosudarstvennogo tehničeskogo universiteta. seriâ: upravlenie, vyčislitelʹnaâ tehnika i informatika
Language(s) - English
Resource type - Journals
eISSN - 2224-9761
pISSN - 2072-9502
DOI - 10.24143/2072-9502-2019-4-95-105
Subject(s) - computer science , expert system , inference , artificial intelligence , machine learning , fuzzy logic , data mining , object (grammar) , fuzzy rule , sensitivity (control systems) , context (archaeology) , inference engine , rule of inference , fuzzy set , engineering , paleontology , electronic engineering , biology
The article describes a method of morphogenesis of the rulebase of fuzzy conclusion system in the context of counter-expert opinions. One of the difficulties of the morphogenesis of the rule base, which reflects the result of the collective opinion of the expert group, is processing of counter-narrative conclusions, which are expressed as incompatible rules. The methods used to formulate the rule bases of fuzzy inference systems, in the case of counter-predictive opinions of experts, are based on the removal of incompatible rules, depending on the value of their confidence coefficient. Moreover, the methods for identifying the values of confidence coefficients are not described enough, in addition, the removal of the rules leads to the loss of information about the object that the expert group formed. The proposed method of morphogenesis of rule bases in terms of counter-predictive expert opinions based on the results of the interaction of input variables is based on the identification of the confidence coefficient of each of the rules, depending on the number and level of qualification of the experts who proposed it. To evaluate the effectiveness of the proposed method, a numerical experiment was carried out, based on the study of a typical model for assessing the state of an object, which is used in other examples that demonstrate the principles of the fuzzy inference system. To compare the effectiveness, expert information processing methods containing incompatible rules were used. The performance criterion was the model sensitivity indicator. In the framework of the experiment, sensitivity was understood as the number of various values of the output variable depending on the values of the input parameters. As a result of the experiment, it was shown that the fuzzy inference system using the rule base formed using the proposed method has a noticeably wide variety of input values while maintaining the monotonicity of the change in the values of the output variable. The results of the study allow more advanced methods for identifying the state of elements of socio-economic and organizational-technical systems in which there is terminological uncertainty in the description of critical parameters and the incomplete knowledge of experts on a problem area.