讨论了化工过程数据校正中数据分类的作用和现有的算法 ,证明了用投影矩阵法对含有未测回路的流程网络进行数据分类的局限性。提出了一种新的利用关联矩阵中各物料流列向量间的线性相关关系的数据分类算法 :线性相关法。为检测网络的合理设计提供了一些理论指导 The role of data classification and existing algorithms in chemical process data correction are discussed. The limitations of using projection matrix method to classify the process network with untested loops are proved. A new data classification algorithm using the linear correlation between each stream column vector in the correlation matrix is ​​proposed: linear correlation method. It provides some theoretical guidance for testing the reasonable design of the network