published in  Proc. of National Conference on Computing and Information Technology, Thailand, May 9-10, 2012.

A Heuristic Approach to Improve The Decision Tree of Product Acceptance in Hard Drive Manufacturing

Siltepavet, A., Sinthupinyo, S., and Chongstitvatana, P.,
Department of Computer Engineering
Chulalongkorn University

email: prabhas@chula.ac.th

Abstract

This paper is aimed at constructing a model which can improve quality of products in hard drive manufacturing by Decision Tree Learning technique. The output model is a guide line for parameters adjustment in the development-process. It advises the best method to adjust controllable parameters that take the maximum pass-pieces. The analysis of this model is divided into two parts. The first part provides an analysis on a decision tree that can determine the parameters which affect the quality of the work-pieces. In the second part, the adjustability of the parameter is discussed and analyzed whether it can be adjusted. Furthermore, two approaches are illustrated to identify the adjustment methods, i.e. the Maximum-Fail First search, the Minimum-Fail First search then the best method which gains more quantity of pass-pieces will be selected.

Keyword: Decision Tree, Hard drive Manufacturing, Improving Quality of Product.