Title: Evaluating and predicting the quality performance in apparel: an application of data science techniques

Authors: B.R.P.M. Basnayake; A.P. Hewaarachchi; N.V. Chandrasekara

Addresses: Department of Statistics and Computer Science, Faculty of Science, University of Kelaniya, Sri Lanka ' Department of Statistics and Computer Science, Faculty of Science, University of Kelaniya, Sri Lanka ' Department of Statistics and Computer Science, Faculty of Science, University of Kelaniya, Sri Lanka

Abstract: In the apparel industry, the quality of the products is significant for the success of organisations. The first time through (FTT) value is an indicator used to identify the performance of the quality of the garment. This work represents a case study concerning a garment factory in Sri Lanka which exports branded clothing. The main objective is to build models with data science techniques rather than applying traditional statistical techniques to predict the FTT of the products with higher accuracy. Data science techniques: regression tree, feedforward neural network (FFNN) with Levenberg-Marquardt (LM) and scaled conjugate gradient (SCG) algorithms and radial basis neural network were used to predict the quality. The results suggested that the error was lowest in the FFNN with SCG. In this sense, the vital advantages of quality prediction are to reduce the rejected products and unnecessary production costs and to achieve the growth of the company.

Keywords: apparel industry; first time through; FTT; regression tree; neural network; prediction.

DOI: 10.1504/IJBDA.2022.126805

International Journal of Business and Data Analytics, 2022 Vol.2 No.2, pp.171 - 186

Accepted: 25 Jul 2022
Published online: 07 Nov 2022 *

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