A New Support Vector Machine Optimized by Simulated Annealing for Global Optimization

摘要

SA-SVM model was proposed in which parameters were optimized by simulated annealing. Parameter (the kernel function) and C (the error discipline) are the key factors to the precision of SVM. Simulated annealing was used to optimize the key parameters of SVM to make enhancement on the forecasting effect of SVM. By applying this proposed model for several function optimizations, results of which demonstrate the improvement of SA-SVM on the high model accuracy in the optimization searching, and it can overcome the blindness of the model parameters.