基于粗集理论的规则知识获取

粗集理论是处理不完整数据和不精确、不确定性问题的理论．基于粗集理论，分析实到诊断数据，获取诊断规则知识．它将要处理的问题表达为知识表达系统，由条件和决策属性的依赖关系消去冗余条件属性，根据属性取非正常值的数目决定其重要性束决定冗余属性取舍。通过计算各规则的核值获取系统的简化属性核值表，采用规则优化得到简化规则集，在简化属性核值表中取值次数的多少决定规则选取的优先次序．并以武器系统中某设备模块的一组实到故障诊断数据为例，具体说明了诊断规则知识的获取过程。

Rules Knowledge Acquisition Based on Rough Sets Theory

Rough set theory is used to deal with incomplete information, inaccurate and uncertain problem. Through analyzing and processing actual data, the method of fault diagnosis rules acquisition based on rough set theory was presented. The expected problem is expressed as a system for knowledge expression, its redundant condition attributes are expurgated according to the reliant relations of its condition and decision attributes. Whether a redundant attribute will be accepted or rejected depends on the importance that is determined by the amount of its abnormal values in the original data. The rule core table is obtained by calculating, then, a reduction rule set is received by optimization. The priority order of a rule in the reduction rule set is decided by the times where its value appears in the core table. Taken a set of actual test and diagnosis data of a certain model in a weapon system as example, the course of rule knowledge acquisition for actual test and diagnosis was introduced.