Application of SVM for Predicting Tensile Characteristics of Cotton Ring Yarn

摘要

High Volume Instruments (HVI) and Advanced Fibre Information System (AFIS) have revolutionized the concept of cotton fibre testing. Test results on cotton fibres have been used by the researchers to predict the yarn properties using mathematical, statistical and artificial neural network (ANN). This estimation can be extended to Support Vector Machine (SVM) and the accuracy of prediction in case of SVM model is found better in most cases in the present study. The SVM needs less number of training data compare to ANN for getting high degree accuracy. The SVM is basically a classifier where there is no direct numerical output but the tenacity values are converted into some predefined class labels for obtaining output results. This paper describes an application of SVM model for predicting the yarn strength from fibre properties and comparison of this result with the ANN method.