Early warning systems for multi-variety and small batch manufacturing based on active learning

Issue title: Applied Mathematics Related to Nonlinear Problems

Guest editors: Juan L.G. Guirao and Wei Gao

Article type: Research Article

Authors: Li, Quanbao^a | Wei, Fajie^a | Zhou, Shenghan^{b; *}

Affiliations: [a] School of Economics and Management, Beihang University, Beijing, China | [b] School of Reliability and System Engineering, Beihang University, Beijing, China

Correspondence: [*] Corresponding author. Shenghan Zhou, School of Reliability and System Engineering, Beihang University, Beijing 100191, China. Tel./Fax: +86 1082317804; E-mail: zhoush@buaa.edu.cn.

Abstract: Multi-variety and small batch production is one type of mainstream production methods. Currently, methods of enterprise risk warning have been abundantly researched by scholars, but the effect of its application to the multi-variety and small batch manufacturing practices is not ideal. In this study, the authors apply the transductive support vector machine and active learning to the study of enterprise risk and early warning methods. The experiment utilizes real-world enterprise data and demonstrates that this method may meet the practical needs of the enterprise risk early warning systems and contributes to solving problems of multi-variety and small batch manufacturing operations.

Keywords: Early warning, risk management, active learning, transductive support vector machine, semi-supervised learning

DOI: 10.3233/JIFS-169345

Journal: Journal of Intelligent & Fuzzy Systems, vol. 33, no. 5, pp. 2945-2952, 2017