Optimal scale selection by integrating uncertainty and cost-sensitive learning in multi-scale decision tables

Optimal scale selection is an important issue in the study of multi-scale decision tables. Most existing optimal scale selection methods have been designed from the perspective of consistency or uncertainty, and cost as well as user requirements or preferences in practical applications has not been considered. It is well known that the uncertainty of decision making in different levels of scale varies in sequential three-way decision models. Furthermore, test cost depends on the scale, and delayed decisions may cause delay cost. In practical applications, both uncertainty and cost are supposed to be considered. Therefore, it is worthwhile to introduce cost-sensitive learning into multi-scale decision tables and select the optimal scale by comprehensively considering uncertainty and cost. In this study, uncertainty is firstly quantified, and a novel cost constitution is defined in sequential three-way decision models. In addition, a multi-scale decision information system based on test cost and delay cost is proposed. Then, to obtain the optimal scale with the minimum uncertainty and cost, an optimal scale selection model is established with the constraint of user requirements. Furthermore, an improved optimal scale selection model considering user preferences is proposed by introducing the ideal solution to resolve conflicts among objectives. Finally, the effectiveness of the optimal scale selection model is verified through experiments, and a comparative experimental analysis demonstrates that the proposed model is more consistent with actual user requirements than existing models.