Title: Classification, feature selection and prediction with Neural-network Taguchi System

Authors: Bharatendra K. Rai

Addresses: Department of Decision and Information Sciences, Charlton College of Business, University of Massachusetts – Dartmouth, 285 Old Westport Road, North Dartmouth, MA 02747-2300, USA

Abstract: Mahalanobis-Taguchi System (MTS) is often compared with artificial neural networks as both methodologies share common application areas. However, the comparison has been strictly limited to latter as a standalone process. Neural networks in a MTS framework, due to availability of a large array of architectures, has potential to lend flexibility needed to deal with a wide variety of application areas. This paper proposes a Neural-network Taguchi System (NTS) approach that incorporates neural networks in a MTS framework and consists of four stages viz., plan, validate, identify, and monitor. The workability of the proposed approach is illustrated using a tool-breakage prediction problem.

Keywords: MTS; Mahalanobis-Taguchi system; multilayer perceptron; NTS; neural networks; Taguchi methods; tool breakage prediction; threshold level; tool failure.

DOI: 10.1504/IJISE.2009.026769

International Journal of Industrial and Systems Engineering, 2009 Vol.4 No.6, pp.645 - 664

Available online: 26 Jun 2009 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article