Title: The use of unclear conclusion in the tasks of forecasting of the durability of corrosive constructions
Authors: Larysa Korotka
Addresses: Department Information Systems, Ukrainian State Chemical Technology University, Dnipro, 49008, Ukraine
Abstract: The borders of each cluster are considered as the carrier of a term set, and the centre of the corresponding cluster is considered as its core. It is assumed that the choice of the type of membership function remains according to the expert in the subject area. The necessary and sufficient volumes of the fuzzy rule base, as well as term sets of linguistic variables are considered. When using Mamdani fuzzy inference and the centroid method, it is possible to obtain the parameters of numerical procedures for the task of forecasting the durability of corrosive structures. As an alternative to neural networks in solving approximation tasks, it is proposed to use a generalised algorithm based on the theory of fuzzy sets. The results of numerical experiments make it possible to assert that a fuzzy clusteriser allows to determine rational parameters of numerical procedures for the class of tasks.
Keywords: clustering; fuzzy knowledge base; FKB; fuzzy inference; forecasting of durability; corroding structures.
International Journal of Computing Science and Mathematics, 2021 Vol.14 No.3, pp.263 - 273
Received: 04 Mar 2019
Accepted: 25 Apr 2019
Published online: 23 Dec 2021 *