A novel method is proposed in developing soft sensor based on the nonlinear principal components obtained from kernel principal component analysis (PCA) technique to deal with the nonlinear characteristics of data.

• 摘要提出了基于核函数主元分析(PCA)方法提取变量的特徵信息以有效处理非线性数据，并在此基础上进行软测量建模的方法。