Optimal Scale Combination Selection Based on Cost-Sensitivity and Sequential Three-Way Decision

Optimal scale combination (OSC) selection is a crucial issue in the research of multi-scale decision system (MDS). However, most of the existing information processing procedures for finding the OSC adopt the fine-to-coarse pattern, which contradicts the coarse-to-fine pattern of human cognition. Moreover, most of the existing methods for finding the OSC do not take cost into consideration. To address the issues, the method of OSC selection based on cost-sensitivity and sequential three-way decision (S3WD) is proposed, which minimizes the total cost while preserving the classification ability. Firstly, a new strategy of optimal granularity selection that adopts attribute reduction followed by scale selection for MDS is proposed to remove the redundant attributes. Secondly, a new measurement of scale significance is defined and a forward heuristic optimal scale selection strategy is proposed. Finally, the OSC selection model based on cost-sensitivity and S3WD is proposed to determine the OSC with the minimum total cost while preserving the classification ability. Experimental results demonstrate that the proposed method can significantly reduce the cost of finding the OSC.