IMPROVEMENT USING AN ARTIFICIAL NEURAL NETWORK AND COMPARISON WITH RESPONSE SURFACE TECHNIQUE FOR YIELDING MULTIPLES

Abstract

Cotton slub yarn is frequently utilized in mechanical, physical, and casual conditions as well as denomination. The Department of Polymer Engineering at National Textile University in Faisalabad provided the data for the main objective. Analysis is done using software written in the R programming language. Cotton production is influenced by a number of variables, all of which have a direct impact on process efficiency. In order to maximize numerous yields (elongation, imperfection, strength, coefficient of mass variation, and hairiness), the study aimed to optimize the 100% cotton slub yarn model (slub length, slub thickness, pause length, and linear density).By evaluating a set of quality parameters, such as process efficiency, using two techniques—response-surface methodology (RSM) and artificial neural network (ANN)—and comparing the results using mean square error (MSE), optimization is a means of determining and improving the performance of the built framework. For added accuracy, the mean square error root (RMSE) and coefficients of determination () are employed. But in every category, the ANN has continuously outperformed the RSM. With an RMSE of 0.229, the final ANN model that was chosen was able to predict all five output parameters at once.