Title: A demonstration of the effects of the number of fuzzy states on a control system
Authors: Maged N. Kamel Boulos
Addresses: The Alexander Graham Bell Centre for Digital Health, Moray College UHI, Elgin, IV30 1JJ, UK
Abstract: In this paper, we describe a fuzzy logic demonstrator that uses heart rate and skin impedance values as inputs to predict a hypoglycaemia index. The demonstrator was created using JFS v2.00 fuzzy logic freeware and is freely available on the author's website at http://healthcybermap.org/FL/index.htm. We investigate the effects of varying the number of fuzzy states (adjectives or membership functions) in our demonstrator. Iterative experimentation and tuning are always necessary to determine the ideal number and configuration of fuzzy sets for solving a particular clinical problem. The exercise presented in this paper is meant for educational purposes, to provide new students of this important subject with some fundamental insights into the inner working of fuzzy logic systems in their quest to simulate the decision-making process of a human expert.
Keywords: fuzzy logic; fuzzy states; control systems; clinical decision support systems; hypoglycaemia; diabetes mellitus.
DOI: 10.1504/IJBET.2022.123255
International Journal of Biomedical Engineering and Technology, 2022 Vol.39 No.1, pp.93 - 105
Received: 24 Nov 2016
Accepted: 03 Feb 2018
Published online: 07 Jun 2022 *