计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (3): 43-43.

• 学术探讨 • 上一篇    下一篇

微粒群算法在改进多元线性回归上的应用

孙辉 张忠梅 葛寒娟   

  1. 南昌工程学院
  • 收稿日期:2006-05-30 修回日期:1900-01-01 出版日期:2007-01-21 发布日期:2007-01-21
  • 通讯作者: 孙辉

Application of PSO to Improve on Multiple Regression

  • Received:2006-05-30 Revised:1900-01-01 Online:2007-01-21 Published:2007-01-21

摘要: 文献[1]利用带约束的非线性规划,将各种改进的多元线性回归方法——主成分回归,岭回归,稳健回归及约束回归统一在一个非线性规划模型中。本文应用微粒群优化算法(Particle Swarm Optimization,PSO)对其进行求解,实际算例表明,该方法不但可行,而且得出的结果比其它方法及文献[3]的结果与实际符合得更好。

关键词: 多元线性回归, 模型, 非线性规划, 微粒群算法

Abstract: The paper[1] put forward a nonlinear programming model that unified all kind of multiple linear regression such as Principal Components Regression,Ridge Regression,Robust Regression and constrained regression. This paper we get the solution to the model by use of particle swarm optimization(PSO) method. The example shown that the solution got by PSO method is more close to the real situation.

Key words: multiple linear regression, model, nonlinear programming, Particle Swarm Optimization