Sequential multi-class three-way decisions based on cost-sensitive learning

Abstract

The theory of sequential three-way decisions which is in accordance with human cognition is a multi-stage decision-making progress. In fact, sequential strategy can deal with multi-class decision effectively. However, few researchers have focused on the attribute significance under multi-class decision from the perspective of cost-sensitive learning in the theory of sequential three-way decisions. By considering reconstructing the granular structure of sequential process, a new sequential multi-class three-way decision model is proposed. First, two types of decision cost in the proposed model are defined. Next, the attribute significance based on decision cost for specific class is devised. Then, through the TOPSIS method, a progress for calculating attribute sequence is proposed. Finally, the experimental results show that the performance of the proposed model has lower decision cost compared with the sequential multi-class three-way decisions.