基于断点辨别力的粗糙集离散化算法

Continuous attributes discretization algorithm based on cut's discriminability in rough set

提出了基于断点辨别力的粗糙集离散化算法通过分析候选断点与决策类之间的相关性,定义了候选断点对决策类的辨别力,并以此作为断点重要性的度量,实现连续属性的离散化。离散化后的决策系统不改变原有的相容性,而且能最大限度地保留有用信息。采用多组数据对该算法的性能进行了检验,并与其它算法做了对比实验。实验结果表明该算法是有效的。

A new algorithm for continuous attributes discretization based on cut ＇s discriminability was proposed. The cut points have been obtained by analyzing the discriminability of initial candidate points to each decision class. With this algorithm, decision systems after discretization guarantee the initial consistency and can reserve useful information as much as possible. In this paper, a group of data set was applied to test the performance of the algorithm and the experimental result was compared with that of other discretization algorithm. The experimental result shows that the algorithm is effective and keeps a high computing efficiency when the number of candidate cut point increases.