MISO系统的混合核函数LS-SVM建模

Modeling of LS-SVM based on mixtures of kernels for MISO systems

将LS-SVM非线性建模思想应用于锌钡白转窑煅烧过程的MISO系统建模.研究发现,系统选取不同的核函数,对模型的拟合性能和预测(泛化)性能有很大的影响.采用基于混合核函数的LS-SVM建模方法解决上述问题,该方法可使系统具有满意的模型拟合输出,能有效抑制局部核函数所引起的预测输出波动,取得了良好的综合辨识效果.

Least squares support vector machines (LS-SVM) is used to solve the multi-input and single output (MISO) modeling problem of lithopone calcination process in rotary kiln. The modeling research discovers that the kernels of LS-SVM have deep influence on fitting and prediction (generalization) performance. To solve this problem, a new kind of LS-SVM based on mixtures of kernels is proposed. This new method not only makes the system get a satisfied fitting output, but also effectively restrains the fluctuation of the prediction output caused by local kernels.