Generalized multiview sequential three-way decisions based on local partition order product space

Abstract

The hierarchical sequential three-way decision model is a method for addressing complex problem-solving. The existing hierarchical sequential three-way decision models mostly employ multi-view and/or multi-level approaches. However, as the number of views increases and the levels deepen, the model becomes too large to solve problems efficiently. In order to solve this problem, this paper proposes a generalized multiview hierarchical sequential three-way decisions based on local partition order product space model. Specifically, we first use a nested partition sequence to represent a view. Next, the linear order relations between levels within the views are split according to the number of levels to obtain local linear order relations. Then, in the multiple views, the local linear order relations between levels close to each other from different views are combined using Cartesian product operations to construct a generalized local partition order product space. Finally, by integrating the hierarchical sequential three-way decisions, the generalized local partition order product space is transformed into a multiview hierarchical sequential three-way decisions model. Experimental results on multiple datasets demonstrate that the proposed multiview hierarchical sequential three-way decision model achieves better performance compared to the existing models.