Affiliations: School of Economics and Management, Southwest Jiaotong University, Chengdu, 610031, P.R. China, [email protected] | School of Information Science and Technology, Southwest Jiaotong University, Chengdu, 610031, P.R. China, [email protected] | School of Management and Engineering, Nanjing University, State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, 210093, P.R. China, [email protected]
Note: [] This work is supported by the Major Program of National Natural Science Foundation of China (No. 71090402/G1), the National Science Foundation of China (Nos. 61175047, 60873108, 70971062), the Youth Social Science Foundation of the Chinese Education Commission (11YJC630127), the Natural Science Foundation of Jiangsu, China (BK2011564). Address for correspondence: School of Economics and Management, Southwest Jiaotong Univ., Chengdu, 610031, P.R. China
Abstract: By considering the levels of tolerance for errors and the cost of actions in real decision procedure, a new two-stage approach is proposed to solve the multiple-category classification problems with Decision-Theoretic Rough Sets (DTRS). The first stage is to change an m-category classification problem (m > 2) into an m two-category classification problem, and form three types of decision regions: positive region, boundary region and negative region with different states and actions by using DTRS. The positive region makes a decision of acceptance, the negative region makes a decision of rejection, and the boundary region makes a decision of abstaining. The second stage is to choose the best candidate classification in the positive region by using the minimum probability error criterion with Bayesian discriminant analysis approach. A case study of medical diagnosis demonstrates the proposed method.
