Mixed data-driven sequential three-way decision via fusion

In the context of granular computing, sequential three-way decision is a useful tool to triadic thinking, triadic computing and triadic processing from coarser to finer under multilevel and multiview granularity space. In this paper, we mainly explore a novel framework of sequential three-way decision for the fusion of mixed data from the subjective and objective dynamic perspectives. The former focuses on the decision maker's dynamic behavior without considering the time-evolving data, and the latter emphasizes on dealing with dynamic mixed data over time by multi-stage decision-making. We firstly utilize four T-norm operators and kernel-based similarity relations to integrate different types of dynamic data. Then the subjective and objective models of sequential three-way decision are investigated based on decision thresholds, attribute importance and cost reduction. Finally, the comparative experiments are reported to verify that our proposed models can achieve the lower decision cost and the acceptable accuracy. (c) 2021 Elsevier B.V. All rights reserved.

Automated summary

In this paper, a novel framework of sequential three-way decision for the fusion of mixed data from the subjective and objective dynamic perspectives is explored. The proposed models achieve lower decision cost and acceptable accuracy, as demonstrated by comparative experiments.