Application study on nonlinear dynamic FIR modeling using hybrid SVM-PLS method

Hybrid support vector machines and partial least squares (SVM-PLS) method for modeling was proposed and was applied to develop nonlinear dynamic finite impulse response (FIR) models in order to improve the performances of the model. Firstly the theory of support vector regression machines and PLS was briefly described. Secondly the principals and framework of hybrid SVM-PLS method were introduced. This method integrated the merits of both SVM and PLS. Thirdly the steps in developing nonlinear FIR model using hybrid SVM-PLS method were given. Finally the superior performances of the nonlinear FIR model were demonstrated by an application study on a chemical process.