Performance of different membership functions in stress classification with fuzzy logic

Abstract

Stress has become an indispensable part of today's world. Stress can have a very serious negative impact on human health. Knowing the intensity of stress on people is important in order to cope with it. In this research, 4 different Fuzzy Logic (FL) structures were used to classify human stress through sleep. In the established structures, the human stress detection data set in sleep and through sleep obtained from Kaggle was used. In the FL structures created, blood oxygen level and respiratory rate were taken as input and stress classification was made accordingly. Their performance in the classification of sleep stress was evaluated by using different membership functions in 4 different structures. In order to make a fair comparison in the established structures, the FL parameter was determined the same, except for the membership functions. As a result of experimental studies, the F model established with the generalized bell showed more successful results than the models established with other membership functions.