An Improved QPSO Algorithm for Parameters Optimization of LS-SVM

摘要

Aiming at the parameter optimization of least square support vector machine (LS-SVM), an improved quantum-behaved particle swarm optimization (IQPSO) algorithm for LS-SVM parameter selection was proposed. Based on QPSO, the algorithm optimizes particle initializing positions and improves solving speed and precision by sampling and linearizing methods. IQPSO LS-SVM model was test by test functions and was compared with QPSO LS-SVM model. Furthermore, it was applied to thread's amount setting of database server in an agricultural producing system. The results show that the proposed model has greater solving speed and higher precision. It can meet the database's load requirement by thread's amount adjustment in agricultural producing system.

关键词