摘要
在分析粗糙集理论、分层聚类算法和k-means聚类算法的基础上,提出一种基于粗糙集和混合聚类法的决策表约简算法,该算法首先是使用基于分层聚类的k-means混合聚类法离散化决策表中的连续属性,然后利用粗糙集理论对离散后的决策表进行属性约简,得到决策规则集,并通过在铁路客运量预测系统中的应用验证了算法的可行性和有效性.
Based on rough set and hybrid clustering method, rough set theory, hierarchical clustering algorithm are analyzed, and an algorithm for decision table reduction is proposed. Firstly, the algorithm discretizes continuous attribute in decision table using hierarchical k-means hybrid clustering algorithm. Then the attribute of the discretization decision table is reduced by using rough set theory, and decision rule gather is extracted. Lastly the feasibility and validity of algorithm is tested though applying the algorithm in the railway passenger volume forecast system.
出处
《大连交通大学学报》
CAS
2008年第3期86-90,共5页
Journal of Dalian Jiaotong University
基金
教育部科学技术研究资助项目(204032)
关键词
粗糙集
混合聚类法
离散化
属性约简
规则提取
rough set theory
hybrid clustering method
discretization
attribute reduction
extracting rules
