挖掘数据中实时数据预处理技术的应用  

Application of Real-time Data Pre-processing Technology in Data Mining

作　　者：陈鑫[1] 赵霞[1] 

机构地区：[1]上海交通大学自动化系

出　　处：《微型电脑应用》2010年第9期34-35,41,共3页Microcomputer Applications

摘　　要：流程工业中数据挖掘结果,是以所需参数的完整性和准确性为前提的。本文综合流程工业实际情况,设计出一种基于支持向量机的实时数据预处理混合算法,通过该预处理算法,可以处理因测量仪表故障导致的数据缺失和外部干扰引起的实时数据采集误差,提高进入数据挖掘系统的数据的质量,并可对仪表故障实时诊断与报警,同时给出该实时数据预处理算法的实施步骤。基于该预处理算法设计开发的实时数据预处理系统已成功应用于某大型化工企业,获得了满意的效果。Process Industry Data mining results based on the accurate measurement of the required parameters, this paper presents a SVM-based hybrid algorithm of real-time data pre-processing. Through this algorithm, the accuracy of data is greatly improved which will access to data mining system. Design thinking and implementation details of this algorithm are also presented here, and the real-time data pre-processing system based on this algorithm is successfully applied to a large-scale chemical industry with satisfactory results.

关 键 词：流程工业 支持向量机 实时数据预处理 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]