Open AccessMETHOD FOR OBTAINING THE PARAMETERS OF MEMBERSHIP FUNCTIONS OF FUZZY SETS BASED ON REAL DATA FOR AUTOMATED INFORMATION PROCESSING SYSTEMS
Author(s) -
E. E. Bisyanov,
Andrey Gutnik
Publication year - 2019
Publication title -
vestnik dagestanskogo gosudarstvennogo tehničeskogo universiteta. tehničeskie nauki
Language(s) - English
Resource type - Journals
eISSN - 2542-095X
pISSN - 2073-6185
DOI - 10.21822/2073-6185-2019-46-3-79-86
Subject(s) - membership function , data mining , defuzzification , fuzzy logic , fuzzy set , fuzzy number , computer science , fuzzy classification , object (grammar) , gaussian function , set (abstract data type) , type 2 fuzzy sets and systems , kernel (algebra) , mathematics , artificial intelligence , gaussian , physics , quantum mechanics , combinatorics , programming language
Objectives Development of a method for selecting the type of accessory function and obtaining its parameters to allow subjective personal influences in automated information processing to be excluded. Method . Existing methods for constructing membership functions were analysed. The research was based on the methods of fuzzy logic and data analysis. Results . A method for obtaining the parameters of membership functions of fuzzy sets using real data is suggested. It is proposed to use the data obtained from the object under study to determine the kernel of the fuzzy number, as well as derive theoretical information about the object for the carrier. Triangular, trapezoidal, bell-shaped and Gaussian membership functions are considered. The appearance of the membership function can be defined using the criterion of the relations of the kernel to the carrier of a fuzzy set. The results of calculations for obtaining the membership functions based on data on the power consumption of electric motors of different types are given. Conclusion . The developed method can be used both in decision support systems as well as in automated systems for controlling technological processes. If necessary, the values of the criterion proposed in the article can be revised to take the values included in the set of measured real data into account or to simplify the procedure of automated processing. Further research will use the described method to obtain the terms of linguistic variables.