Predicting the Hairiness of Ring Spinning Polyester/Cotton Yarn Using Multiple Regression and Artificial Neural Network Approaches

Two modeling methods are used to predict the hairiness of polyester/cotton yarn. Excellent agreement is obtained between these two approaches. A neural network model provides quantitative prediction of yarn hairiness. A multiple regression model is very easy to use, by fitting to historical data gathered from experiments. In conclusion, ANN and multiple regression models both have given satisfactory predictions. However, the predictions of ANN gave reliable results than that of multiple regression models. Since the prediction capacity of multiple regression model is also obtained as satisfactory, it can also be used for hairiness prediction of polyester/cotton blended yarns because of its simplicity and non-complex structure.