Fuzzy linear regression model on mulberry silk cocoon characteristics

Abstract

Purpose This study aims to propose a fuzzy linear regression (FLR) model to deal with the vagueness or fuzziness of the underlying relationship between silk cocoon and yarn quality. Design/methodology/approach Shell ratio percentage, defective cocoon percentage and cocoon volume are considered as significant independent variables to predict the quality of silk cocoons. Input and output parameters of the FLR model are considered as non-fuzzy, but the underlying relationship between the variables is assumed to be fuzzy. Findings The fuzzy regression model shows its superiority against conventional multiple linear regression model for estimation of silk cocoon characteristics. It is inferred that the fuzziness in underlying relationship between the parameters can be handled efficiently by FLR model. Originality/value A rigorous experimental work has been carried out on 40 lots of mulberry silk cocoons to generate real-world data set to characterize silk cocoons’ quality in a fuzzy environment.

The Use of an Internet of Things Data Management System Using Data Mining Association Algorithm in an E-Commerce Platform

“…The advantages of using a linear regression model for analysis are a short time for solution, simple form, and relatively easy to establish model. However, it can only fit the linear relationship among variables but not the nonlinear one (Zhu et al, 2021;Kar et al, 2019;Chung et al, 2022).…”