铝电解监控系统重复记录辨识与噪音识别算法

来源期刊：中南大学学报(自然科学版)2004年第6期

论文作者：陈湘涛 张建中 李劼 刘业翔

文章页码：915 - 919

关键词：铝;电解;数据预处理;属性集;重复记录;噪音识别;算法

Key words：aluminum; electrolysis; data pretreatment; attribute sets; repeat record; noise identification; algorithm

摘    要：在新型的铝电解网络监控系统中,需要对大量的监控数据构建数据仓库,为此,针对其重复记录的数据预处理问题进行了研究。在属性集的基础上,提出属性重复量、记录重复量等定义,并给出重复记录辨识的一般算法IRA及其推广算法GIRA;对重复记录集中的噪音处理,提出一种基于距离的噪音识别算法INA;以实例说明了算法的具体实现过程。

Abstract: Problems of data pretreatment about repeat records in the data ware house were studied and the definitions of attribute repeating measuring, record repeatingmeasuring etc. were brought forward based on the attribute sets. The generic algorithm named IRA and itsgeneralized algorithmGIRAwas described for identifying the repeat records. To dispose the noises of repeat record sets, an algorithmof INAbased on distance was put forward. And examples to show that the implementation of these algorithms were indicated in detail.