基于孤立点分析的设备维护案例不一致性辨识研究  

Research on the identification of outlier analysis’ plant maintenance case inconsistency

作　　者：邵继元[1] 于德介[1] 刘坚[1] 

机构地区：[1]湖南大学机械与汽车工程学院,湖南长沙410082

出　　处：《机械》2007年第3期21-23,共3页Machinery

摘　　要：随着CBR系统在设备维护知识管理中的广泛应用,案例库的规模在不断的膨胀。不一致的案例增多,降低了案例库的质量,影响了CBR系统的性能。为此,把孤立点分析算法应用于设备维护案例不一致性的辨识,并对一组实际的风机机组案例库片断进行孤立点分析,找出了不一致案例。此方法与现有的不一致案例检测方法相比,不依赖已有的特征值数据模型,不需要事先建立数据规则,因此具有通用性。As the broad application of CBR system in plant maintenance knowledge management, the size of case base is expanding, and the inconsistent cases are augmenting. As a result, the case bases increases, and the capability of the CBR system is affected. So, a technology of plant maintenance case inconsistency identification on Outlier analysis is provided, and in this way, an inconsistency case is found out from a set of blower fan team＇s case bases. Comparing with the existing inconsistency case measuring methods, this technology, independent of the existing eigenvalue data model, no use for setting data rules in advance, is currency.

关 键 词：基于案例的推理 案例库维护 孤立点分析 不一致性 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]