Scholarlink

<p>Multi-scale set-valued decision tables (MSDTs) are a special type of set-valued information system, characterized by the fact that each attribute of an object can have different values across various scales. The optimal scale selection (OSS) in MSDTs has emerged as a critical issue in knowledge discovery of granular computing. However, existing studies have primarily focused on static data environments, and there remains a lack of research on the OSS problem in dynamic data environments, which are commonly encountered in practical applications. Inspirating from the three-way decisions (3WD), this paper proposes methods of incremental approaches for OSS in MSDTs with increments of attributes and attribute values respectively. Firstly, we introduce a 3WD model in MSDT. Subsequently, utilizing the incremental learning approach, we explore the dynamic mechanisms of the dominance relation and the partition of 3WD regions in MSDT with increments of attributes or attribute values. Furthermore, we develop an updating method for the uncertainty degree of scale transformation in MSDT. Based on the aforementioned studies, we finally present two incremental algorithms i.e., OSS-MSDT-AMV and OSS-MSDT-AMA for the OSS in MSDT with increments of attributes and attribute values respectively. Comparative experiments demonstrate that the proposed incremental algorithms can select the optimal scale of MSDT in dynamic data environments with increments of attribute and attribute values efficiently.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Accelerated optimal scale selection in dynamic multi-scale set-valued decision tables

  • 期刊论文

  • 发表日期：

    2025年7月6日

查看全文
  • Ruihua Wang,
  • Yuanjian Zhang,
  • Yuandong Huang,
  • Jianfeng Xu

摘要

Multi-scale set-valued decision tables (MSDTs) are a special type of set-valued information system, characterized by the fact that each attribute of an object can have different values across various scales. The optimal scale selection (OSS) in MSDTs has emerged as a critical issue in knowledge discovery of granular computing. However, existing studies have primarily focused on static data environments, and there remains a lack of research on the OSS problem in dynamic data environments, which are commonly encountered in practical applications. Inspirating from the three-way decisions (3WD), this paper proposes methods of incremental approaches for OSS in MSDTs with increments of attributes and attribute values respectively. Firstly, we introduce a 3WD model in MSDT. Subsequently, utilizing the incremental learning approach, we explore the dynamic mechanisms of the dominance relation and the partition of 3WD regions in MSDT with increments of attributes or attribute values. Furthermore, we develop an updating method for the uncertainty degree of scale transformation in MSDT. Based on the aforementioned studies, we finally present two incremental algorithms i.e., OSS-MSDT-AMV and OSS-MSDT-AMA for the OSS in MSDT with increments of attributes and attribute values respectively. Comparative experiments demonstrate that the proposed incremental algorithms can select the optimal scale of MSDT in dynamic data environments with increments of attribute and attribute values efficiently.