基于变网格聚类的离群数据的研究

Study on Outliers Data Based on Variable Grid Clustering

随着计算机技术的迅速发展,计算机在过程控制中起着越来越重要的作用,在集散控制系统中,上位机通过PCU(Process ControlUnit)传送来的数据进行分析处理而后对PCU实现参数整定、报警与管理。因此这一领域的数据挖掘方法研究具有十分重要的意义。文章将根据这一领域的特殊情况,提出一种基于变网格的聚类分析算法分析过程数据库中的离群数据,从而实现系统的预测报警功能。

With the development of computer technologies, computers become more and more important during control system processes. Host computer analysis the database which is transmitted from PCU（Process Control U- nit） and then realize parameters tune ,alarms and giving something administer on PCU in distributed control sys- tem. So the importan of database mining algorithms is unambiguous in process control fields. Some special case can be considered in the fields. , In this paper one algorithm based on variable grid for outliers data are put for- ward to analysis outliers data in process database. Finially the capabiltity of prediction and alarm system can be improve.