A study on siro, solo, compact, and conventional ring-spun yarns. Part III: modeling fiber migration using modular adaptive neuro-fuzzy inference system

Abstract

The importance of fiber migration in spun yarns as a means of securing cohesion and strength has been emphasized in the literature. However, analyzing migration behavior of fibers is a time-consuming and tedious task. A three-stage hybrid model was developed to estimate yarn migratory properties based on some physical and mechanical properties of spun yarns. Achieving the objectives of this research, general physical, mechanical, and structural properties of spun yarns together with existing standards were thoroughly studied. At the first stage, using stepwise regression analysis, key variables were selected. At the second stage, data-set was clustered into subpopulations by means of K-means in order to decrease effects of noise, rebate complexity of the patterns, and develop a modular model. At the third stage, using adaptive neuro-fuzzy inference system, the target value was predicted. Finally, evaluation of the proposed model was carried out by applying it on the test set.