摘要
作为“数据挖掘”的基本技术之一的多元统计技术在数据的降维、分析方面有着广泛的应用。鉴于MPCA(Multi-wayPrincipalComponentAnalysis)在处理批生产数据方面的优越性,以链霉素发酵生产过程为例,用MPCA方法和生产历史数据建立了统计模型,并将模型用来对生产进行实时监视、故障诊断和效果评价。仿真结果证明了该方法有很好的应用效果。
A large amount of data is collected by computer control system in the batch process industry.In order to extract valuable information from t he historical database,as the key te chnology of Data Mining,MPCA(Multi -way Principal Component Analysis)has been shown effective.This method is capable of monitoring the new runs and detecting the occurrence of obse rvable upsets by constructing a know ledgebased model.The effectiveness of MPCA is illustrated through a simula tion of fermentation process to show potential application.
出处
《自动化与仪表》
2002年第3期29-31,共3页
Automation & Instrumentation
