东北大学学报:自然科学版 ›› 2014, Vol. 35 ›› Issue (2): 170-174.DOI: 10.12068/j.issn.1005-3026.2014.02.005

• 信息与控制 • 上一篇    下一篇

基于黑洞算法的LSSVM的参数优化

王通1,高宪文1,蒋子健2   

  1. (1.东北大学 信息科学与工程学院, 辽宁 沈阳110819; 2.沈阳工业大学 电气工程学院, 辽宁 沈阳110870)
  • 收稿日期:2013-05-27 修回日期:2013-05-27 出版日期:2014-02-15 发布日期:2013-11-22
  • 通讯作者: 王通
  • 作者简介:王通(1976-),男,辽宁沈阳人,东北大学博士研究生;高宪文(1955-),男,辽宁沈阳人,东北大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金重点资助项目(61034005).

Parameters Optimizing of LSSVM Based on Black Hole Algorithm

WANG Tong1, GAO Xianwen1, JIANG Zijian2   

  1. 1. School of Information Science & Engineering, Northeastern University, Shenyang 110819, China; 2. School of Electrical Engineering, Shenyang University of Technology, Shenyang 110870, China.
  • Received:2013-05-27 Revised:2013-05-27 Online:2014-02-15 Published:2013-11-22
  • Contact: WANG Tong
  • About author:-
  • Supported by:
    -

摘要: 采用黑洞(BH)算法对最小二乘支持向量机(LSSVM)的惩罚系数C及径向基核函数参数σ进行搜索优化,提高LSSVM的预测性能.黑洞算法模拟自然界黑洞,吸引一定范围内的星体向其运行并吸收它们;算法在运行过程中,始终保持黑洞为最优解,通过星体的运行搜索整个空间.通过基于黑洞算法的LSSVM和基于粒子群(PSO)算法的LSSVM实现对二维函数的预测,并对二者进行了仿真研究.仿真结果证实,黑洞算法可以更好地实现LSSVM参数的优化搜索,且基于黑洞算法的LSSVM方法具有更高的预测精度.

关键词: 黑洞算法, 最小二乘支持向量机, 参数搜索, 粒子群优化, 二维函数

Abstract: Black hole(BH) algorithm is used to search the optimal parameters of least squares support vector machine(LSSVM). Black hole phenomenon in the nature is simulated in BH algorithm, which has the ability to attract the stars moving to it and absorb them in a certain range. During BH algorithm running the entire space is searched through the moving stars and the black hole is considered the optimal solution.In the simulation experiment the twodimensional function is predicted by BHLSSVM and PSOLSSVM respectively. The simulation results show that BH algorithm has excellent capability to search optimal parameters of LSSVM, and BHLSSVM has better capability and higher predictive accuracy.

Key words: black hole algorithm, least squares support vector machine, parameter selection, particle swarm optimization, twodimensional function

中图分类号: