Adaptive Neuro-fuzzy Inference System with Subtractive Clustering: A Model to Predict Fiber and Yarn Relationship

Abstract

The spinning process is an important process in the Textile Industry. The yarn (output) coming out of the spinning process has a unique relationship with the input fibers. The input and output properties show a non-linear relationship between them. Many techniques, such as multiple regression, Artificial Neural Network (ANN) model, etc, have been used in an attempt to develop the said relation. In this paper, the Adaptive Neuro-Fuzzy Inference System (ANFIS), combined with subtractive clustering, is used to predict yarn properties. The ANFIS technique is evaluated for different datasets and its performance is compared with the ANN model.