Abstract:
As for the problem of softsensor modeling estimation precision in chemical processing, a method of combination SVM softsensor modeling is proposed based on a theory of multiple knowledge bases, Supervised Locality Preserving Projection(SLPP). The dimension of input data is reduced by SLPP between clusters, and then different transformed matrices and different multiple knowledge bases are gotten for every category. At last, a combinationmodel is constructed by support vector machine adaptively. Applied this method to a softsensor modeling of bisphenol A content, the simulation results show that every submodel weighted is more reasonable as compared with traditional multimodel, the estimated accuracy of model is improved, and the generalization ability is better.