基于粗糙集的油液不完备多值系统知识发现

Knowledge Discovery from Incomplete and Multi-valued Oil System Based on Rough Set

由于数据量小,油液有效知识的发现,特别是从油液不完备信息系统挖掘决策规则,一直是其有效应用和发展的瓶颈。考虑到分析铁谱对磨损模式的识别贡献最大,结合属性的磨损识别权重和摩擦学系统特点,提出了基于粗糙集的分析铁谱不完备多值系统知识发现模型。通过对部分信息缺省或难以确定的某汽车公司冲压机分析铁谱决策系统进行处理,推导出了关于磨损状态的最优决策规则,说明了粗糙集对不完备油液信息处理有效。

Due to small data, discovery of oil effective knowledge, especially decision rule mined from oil incomplete information system, is always the choke point of effective application and development of oil analysis. Considering the most contribution of analytical ferrography to the recognition of wear mode, knowledge discovery model of incomplete and muhivalued analytical ferrography system based on rough set theory was brought forward combining its attributes weight in the recognition of wear and tribology system traits. Through application to analytical ferrography decision system of stamping machine in an auto company, in which part data is missing or difficult to ascertain, the optimal decision rules about wear state were inferred. It is shown that the good effect of rough set on incomplete oil information processing.