A Teacher-Cost-Sensitive Decision-Theoretic Rough Set Model
Existing DTRS models use two states (success, fail) and three actions (accept, defer, reject) to describe the decision procedure. However, deferment provides a compromise instead of a solution. In this paper, we replace this action with consult which stands for consulting a teacher for correct classification. Naturally, the new action involves the cost of teacher, which is already known in many applications. Through computing the thresholds \(\alpha \) and \(\beta \) with the misclassification cost and the teacher cost from the decision system, the positive, the boundary, and the negative regions are obtained. They correspond to positive rules for acceptance, boundary rules for consulting a teacher, and negative rules for rejection, respectively. We compare the new model with the Pawlak model and previous DTRS model through an example. This study indicates a new research direction of DTRS.
KeywordsDecision-theoretic rough sets Delay cost Misclassification cost Teacher cost Three-way decision
This work is in part supported by National Science Foundation of China under Grant No. 61379089.
- 3.Zhang, H.R., Min, F.: Three-way recommender systems based on random forests. Knowledge-Based Systems (2015). doi: 10.1016/j.knosys.2015.06.019
- 4.Yu, H., Wang, Y., Jiao, P.: A three-way decisions approach to density-based overlapping clustering. In: Peters, J.F., Skowron, A., Li, T., Yang, Y., Yao, J.T., Nguyen, H.S. (eds.) Transactions on Rough Sets XVIII. LNCS, vol. 8449, pp. 92–109. Springer, Heidelberg (2014) Google Scholar
- 10.Yi, J.X., Lin, S.: A simulated annealing algorithm for learning thresholds in three-way decision-theoretic rough set model. J. Chin. Comput. Syst. 11, 2603–2606 (2013). (in chinese)Google Scholar
- 11.Turney, P.D.: Types of cost in inductive concept learning. In: Proceedings of the Workshop on Cost-Sensitive Learning at the 17th ICML, pp. 1–7 (2000)Google Scholar
- 13.Yao, Y.Y., Zhou, B.: Micro and macro evaluation of classification rules. In: 7th IEEE International Conference on Cognitive Informatics, ICCI 2008, pp. 441–448 (2008)Google Scholar