基于支持向量机的软测量技术及其应用

支持向量机(SVM)是一种基于结构风险最小化原理,具有很好推广性能的学习算法。讨论了基于最小二乘支持向量机(LS-SVM)的软测量数据建模原理和方法,并将其应用在汽车排放的氮氧化合物NOX软测量中。通过与基于神经网络的软测量方法进行比较,结果显示出SVM的明显的优势,特别是对小样本、非线性、高维数一类软测量问题的建模,提供了一种有效的途径。

Soft-measurement technique and applications based on support vector machine

The support vector machine (SVM) is an algorithm based on structure risk minimizing principle and high generalization ability.The soft-sensing technique based on least squares-support vector machine(LS-SVM) is presented,and it has been successfully applied to the soft- measurement of NO_X from bus.The result shows that the method based on LS-SVM has obvious advantages than neural network method.Moreover,LS-SVM offers a kind of effective way for soft-sensing model of small sample space,non-linearity,high dimensions.