粗集理论中基于规则及其参数的分类识别

粗集理论的提出为获取if...then...形式的规则提供了重要方法,但是现有的方法一般没有考虑决策表中属性的权重信息。此外,在利用规则进行分类识别时,还可能用到其他方面的信息,如导出规则的等价类、规则前件的长度等。因此,笔者从规则的支持量、精度、适用度、权重、导出规则的等价类和前件长度几个参数对一条规则进行全面描述。利用规则及其参数,提出了几种分类识别的策略。

Classification Based on Rules and their Parameters in Rough Set Theory

Rough set theory has been proved to be a useful tool for rule induction.But ,most of the methods did not consider the weight of the attributes in a decision table.In addition,classification based on the induced rules need other information,such as the number of a rule's antecedent ,objects from which the rule is induced.In this paper,six parameters of a rule,i.e.,support,accuracy,coverage,weight ,condition equivalence classes from which the rule is induced,and the number of antecedent are used to describe the corresponding rule.Then,several techniques for rule-based-classification are put forward.