摘要
粗糙集理论能对系统中的冗余信息进行约简,但其处理过程完全基于样本集,样本集的完备性对其处理结果有直接影响。对粗糙集理论及其在故障诊断中属性约简存在的问题进行了分析,通过实例证明了在故障样本集不完备的情况下,利用粗糙集进行的属性约简会由于新故障样本的引入而导致前后约简结果的不一致,从而影响诊断的准确性;指出了该问题产生的关键原因及解决的办法,并给出了相关的实现算法,以提高系统的故障诊断自适应性。
Rough set theory can reduce the redundant information in the system. But the process of the reduction only based on the sample set. The integrality of the sample set will influence the result directly. Rough set theory and the problem consist in the reduction of the attribute in fault diagnosis are analyzed in this paper. It proved that when the sample set of the fault is incomplete,the result of the attribute reduction will be variant when a new .sample is joined into the sample set, and this will influence the correctness of the diagnosis. Pointed out the pivotal reason and the method that can resolve the problem, and the resolved algorithm is also put forward. It will enhance the ability of the self- adaptation of the system.
出处
《计算机技术与发展》
2008年第1期132-135,共4页
Computer Technology and Development
基金
湖北省教育重点项目(D200614013)
关键词
粗糙集
故障诊断
区分函数
约简
rough set
fault diagnosis
discernibility matrix
reduction