基于LPP-kNN方法的间歇过程故障监视

摘要

In order to address the high dimensionality and multiple conditions of batch process data,a method of LPP-kNN is proposed in this article.Firstly,this method is based on locality preserving projection (LPP) which can extract adaptive transformation matrix of the Vidor High modal batch data to form a new modeling data.Then,standardization of local neighborhood(LNS) is processed to overcome the data charac-ter of multiple conditions.Meanwhile,k-nearest neighbor(kNN) is applied for fault detection with construc-ting statistical indicators.Finally,a variety of improved kNN algorithms are applied in semiconductor indus-try examples and the effectiveness of the proposed method has been verified by comparing.%针对批次过程数据具有高维、非线性及多模态等特性,提出一种自适应LPP-kNN的过程监视方法.利用局部保持映射算法(LPP)提取高维多模态批次数据的自适应变换矩阵构成新的建模数据.采用局部近邻标准化方法(LNS)进行标准化,并利用kNN算法构造统计监测指标.最后,通过在半导体工业实例中的应用验证了所提算法的有效性.