Work

Vector-based rough sets

Year: 2025

Type: article

Abstract: Based on the fact that it is possible to represent approximations in probabilistic rough sets using inner products, we propose vector-based rough set models, where a single vector is treated as a granule. This is achieved by representing each granule, decision class, and approximation as vectors. The proposed models generalize the probabilistic rough set models and inherit properties from these models. Additionally, the geometric properties of approximations within the vector-based framework are discussed. While our previous results required linear independence of the vectors representing decision classes, we demonstrate that this restriction is no longer necessary. Furthermore, we introduce a method for class determination within neural networks using vector-based rough set models, which provides a more reliable class selection.

Source: Information Sciences

DOI: 10.1016/j.ins.2025.122331

Authors: Hajime Okawa, Yasuo Kudo, Tetsuya Murai

Institutions: Muroran Institute of Technology, Chitose Institute of Science and Technology

Language: English

Topic: Rough Sets and Fuzzy Logic

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $3,200

Funder: Japan Science and Technology Agency