Improving the Solution of Least Squares Support Vector Machines with Application to a Blast Furnace System

[摘要] The solution of least squares support vector machines (LS-SVMs) is characterized by a specific linearsystem, that is, a saddle point system. Approaches for its numerical solutions such as conjugatemethods Sykens and Vandewalle (1999) and null space methods Chu et al. (2005) have been proposed. To speed up the solution of LS-SVM, thispaper employs the minimal residual (MINRES) method to solve the above saddle point system directly. Theoretical analysis indicates that the MINRES method is more efficient than the conjugate gradientmethod and the null space method for solving the saddle point system. Experiments on benchmark datasets show that compared with mainstream algorithms for LS-SVM, the proposed approach significantlyreduces the training time and keeps comparable accuracy. To heel, the LS-SVM based on MINRESmethod is used to track a practical problem originated from blast furnace iron-making process: changingtrend prediction of silicon content in hot metal. The MINRES method-based LS-SVM can effectivelyperform feature reduction and model selection simultaneously, so it is a practical tool for the silicontrend prediction task.

[学科分类] 应用数学