论文部分内容阅读
由已知数据中产生的粗糙决策规则往往具有不确定性 ,需要适当的不确定性量度。借鉴变精度粗糙集理论的思想 ,采用基于信息熵的方法构造了两个新的粗糙决策规则不确定性量度函数。它们不仅可以兼顾由划分的粒度引起的规则不确定性的两个方面 ,即不一致性和随机性 ,还考虑了数据中的噪声对规则一致性的影响。因此 ,它们对一类“几乎一致性规则”具有一定的保护作用。通过举例分析 ,说明它们更适于评价从有噪声数据中提取的粗糙决策规则。
Rough decision rules generated from known data are often uncertain and require an appropriate measure of uncertainty. Drawing on the idea of variable precision rough set theory, we construct two new measurement functions based on information entropy. They not only take into account the two aspects of the rule uncertainty caused by the granularity of the division, namely the inconsistency and randomness, but also the influence of the noise in the data on the consistency of rules. Therefore, they have a certain protective effect on a kind of “almost consistent rules”. By example analysis, they are more suitable for evaluating the rough decision rules extracted from noisy data.