Authors: Hao Shao; Rui Xu; Feng Tao
Addresses: School of WTO Research and Education, Shanghai University of International Business and Economics, Shanghai, 201620, China ' School of Business, Hohai University, Nanjing, Jiangsu, 210098, China ' School of Business, East China University of Science and Technology, Shanghai, 200237, China
Abstract: Nowadays, rule mining technologies are widely applied in various areas including management science and operations research, with the objective to find the underlying pattern and tendency as soon as possible in a large dataset as well as to reduce the cost. However, in many real applications, we are often confronted with incomplete datasets and consequently it is not able for us to induce the perfect decision list. To alleviate this problem, we try to incorporate with partial decision lists even the data is incomplete. This method can enable us to respond quickly to real changes of the world. Although partial decision lists shed light on this promising research aspect, one problem is that, to the best of our knowledge, there exists no canonical form of partial decision lists. In this paper we try to provide the canonical form in order to facilitate researchers who are interested in developing new methods for rule mining with incomplete datasets.
Keywords: rule mining; partial classification; partial decision lists; canonical form; incomplete data.
International Journal of Industrial and Systems Engineering, 2015 Vol.19 No.3, pp.422 - 432
Available online: 11 Mar 2015 *Full-text access for editors Access for subscribers Purchase this article Comment on this article