中国民航大学学报 ›› 2021, Vol. 39 ›› Issue (1): 59-64.

• 基础科学 • 上一篇    

基于变精度优粗集模型的一致性规则挖掘算法

陶志,胡柳,王丹   

  1. (中国民航大学理学院,天津300300)
  • 出版日期:2021-02-25 发布日期:2021-02-16
  • 作者简介:陶志(1963—),男,辽宁沈阳人,教授,博士,研究方向为复杂系统建模与优化、智能化建模与软计算方法、粗糙集理论及其应用.
  • 基金资助:
    国家自然科学基金项目(60672178)

Consistency rule mining algorithm based on variable precision dominance relation rough set model#br#

TAO Zhi, HU Liu, WANG Dan   

  1. (College of Science, CAUC, Tianjin 300300, China)
  • Online:2021-02-25 Published:2021-02-16

摘要: 为了在含有噪声数据的优势关系系统中进行数据挖掘,提出以修正的知识粒度作为改进粗糙规则集的不确定性度量,并以此为启发式信息进行属性的相对约简,再依据变精度优粗集所给的参数值导出一致性规则集,并对参数值影响规则集的程度进行灵敏度分析。该算法所得到的一致性规则集具备较高的准确度和覆盖度,同时又可排除由数据噪声所引起的不一致性,从而保证了预测和分类的准确性和合理性。通过实例证明该算法是一种规则获取的有效方法。

关键词: 变精度优粗集, 一致性规则集, 相对粗糙规则集, 相对约简, 阈值平稳区间

Abstract: In order to conduct data mining in dominance-based information system with noise data, the modified knowledge granularity is proposed as the uncertainty measure for the improved rough rule set, and the relative reduction of heuristic information attributes is obtained. Then, the consistency rule set is derived basing on the parameter values provided by variable precision dominance relation rough set, analyzing the influential sensitivity degree of parameter values on rule set. Obviously, the consistency rule set obtained by the algorithm has higher certainty and coverage, the inconsistency caused by data noise is also eliminated, thus ensuring the accuracy and rationality of prediction and classification. Instance analysis proves that this algorithm is an effective method for rule acquisition.

Key words: variable precision dominance rough set, consistency rule set, relative rough rule set, relative reduction, threshold
stationary interval

中图分类号: