利用数据挖掘技术对设备监测数据进行分析,可以建立较准确的故障诊断及预警模型,但随着故障数据库的扩大,如何利用新增数据进行快速诊断成为急需解决的问题。针对上述问题,提出了加权关联规则增量更新模型,该模型直接对新增数据进行频繁项集挖掘,在一定程度上缩减了矩阵规模。通过算例证明了其挖掘结果的准确率明显优于经典的增量模型-FUP。%The fault diagnosis and warning model could be built more accurate with equipment monitoring data analyzed by the data mining technology. But, as the size of fault database scale up, the problem of how to use the incremental data for rapid diagnosis needs to solve urgently. Thus, an incremental updating model based on the matrix-count algorithm is proposed to solve this problem. The model can find the large itemsets in the incremental database directly and the size of matrix is reduced. The example proves that the model is more accuracy than the classic incremental updating model-FUP.
展开▼