摘要
根据决策规则在实际应用中的匹配思想,对数值型一致决策表提出了一种基于模糊聚类方法的决策表约简方法。该方法在保持决策表一致性的前提下,通过冗余度大小的调整,对具有相同决策的对象进行压缩,从而达到对决策表约简的目的。基于计算相关规则的直接分类法验证结果显示,它是有效和可行的。
According to match thought of the decision rules in actual application, a reduction method of decision table based on fuzzy clusterings is presented for consistent decision of keeping consistence of dicision table, the method compresses the universe of decision table by adjusting redundancy degree. Direct classification method based on computing related regular shows its validity.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第5期151-152,162,共3页
Computer Applications and Software
关键词
聚类
相似度
决策表
约简
Clustering Similar degree Decision table Reduction
