Inferential estimation of biopolymer (polyester) quality using bootstrap re-sampling neural network technique

Abstract

Nowadays, biopolymer has been actively used in two important areas in our daily activities; packaging and medical devices. The growing importance of biopolymer has triggered researchers to focus on this matter. One of the important criteria in production of biopolymer is the quality of the product itself. The high quality product is absolutely desirable. Therefore, a method of controlling biopolymer quality is certainly indispensible in this matter. Medical devices certainly demand a high quality biopolymer as these devices always get along with strict specifications in their production. Biopolymerization furthermore is a very nonlinear process which requires a powerful tool to tackle the nonlinearity of the process. Neural network is apparently a powerful tool especially in modeling nonlinear and intricate process. Nevertheless, single network may face problem such as lack generalization capability which can lead to poor performance of the model. Hence, a good alteration to the network is essential to extenuate the problem. Bootstrap re-sampling method is one way to tackle such a job. This work presented a prediction of biopolymer quality using bootstrap re-sampling neural network technique.