Journal of Fibre & Textile Research
Vol. 31, March 2006, pp. 160-169
This paper reports a brief outline of artificial neural network (ANN) and reviews its application in yarn manufacturing process. The use of neural network in predicting process parameters from known and unknown combination of yarn properties has also been investigated. ANN has been found to be very efficient in predicting process parameters when property combinations are taken from actual observed data. However, when the property set is arbitrary, the prediction is poor. The importance of choosing a feasible combination of input parameters has been highlighted.
Keywords: Artificial neural network, Principal component analysis
IPC Code: Int. Cl.8 D02G3/00, G06N3/00