论文部分内容阅读
在提出规则信息量的基础上,分析了可能性规则的定义,并利用规则信息量对条件属性类和决策属性类间的概率因果关系进行了描述,重新定义了可能性规则。同时,进一步提出了能充分利用决策表信息系统中统计信息的规则提取算法LRE。本文还基于规则信息量得到了一种新的规则约简算法,这种算法是在对原有的粗集规则约简定义进行合理推广的基础上建立起来的,能很好地抵抗数据中的噪声。
Based on the quantity of rule information, the definition of probability rule is analyzed, and the probability causality between condition attribute class and decision attribute class is described by rule information, and the rule of probability is redefined. At the same time, the rule extraction algorithm LRE that can make full use of statistical information in decision table information system is further proposed. In this paper, a new rule reduction algorithm based on the amount of rule information is proposed. This algorithm is based on the reasonable promotion of the original definition of rough set rules, which can well resist the data noise.