The application of data mining techniques in quality and reliability prediction and improvement

Abstract

The telecommunications industry is confronted with growing customer expectations and demands for better quality of a lower cost, both of which have added more challenges to the task of quality and reliability prediction and improvement. This research is motivated by a need for the knowledge to support managerial decision making, as one of the wireless devices manufacturers wants to ensure the quality of the product. This research aims to examine these existing business problems and to develop a classification and prediction model, which will give the manufacturer access to information from a range of corporate databases, deemed essential to the manufacturing, activation, and Return Material Authorization databases. The methodology in this thesis is based on a data mining approach, which focuses on the application domain. The results are expected to identify the influential factors that cause the deactivation of the devices, mainly from a product quality point of view.