一种改进的基于知识粒度的增量属性约简算法
Improved Knowledge Granularity-based Incremental Attribute Reduction Algorithm
郑颖春 1郭玲1
作者信息
- 1. 西安科技大学理学院,西安 710054
- 折叠
摘要
不完备混合型决策系统的增量式属性约简问题是近年来研究的热点.针对属性值和属性同时发生变化的情形,给出了一种不完备的混合型决策系统的知识粒度的定义.在对基于知识粒度的增量机制完善的基础上,提出了一种改进的属性值变化且增加属性的增量式属性约简算法.并用UCI上的8个数据集进行仿真实验验证,结果表明,所提的增量式属性约简算法相对于非增量式属性约简算法以及同类型的属性约简算法,在保证分类精度良好的前提下具有较高的约简效率.
Abstract
The problem of incremental attribute reduction for incomplete hybrid decision systems has been a hot topic of research in re-cent years.A definition of knowledge granularity for an incomplete hybrid decision system is given for the case where attribute values and attributes change at the same time.Based on the refinement of the incremental mechanism based on the knowledge granularity,an improved incremental attribute reduction algorithm with attribute value change and increasing attributes is proposed.And the simulation experiments are verified with eight datasets on UCI.The results show that the proposed incremental attribute reduction algorithm has higher reduction efficiency and better classification performance compared with the non-incremental attribute reduction algorithm and the same type of attribute reduction algorithm.
关键词
不完备混合型决策系统/属性约简/知识粒度/增量机制Key words
incomplete hybrid decision system/attribute reduction/knowledge granularity/incremental mechanism引用本文复制引用
出版年
2025
NETL
NSTL
万方数据