计算机科学 ›› 2007, Vol. 34 ›› Issue (6): 174-176.

• • 上一篇    下一篇

基于约束规划的选择性神经网络集成方法

龚安 张敏   

  1. 中国石油大学(华东)计算机与通信工程学院,东营257061
  • 出版日期:2018-11-16 发布日期:2018-11-16

GONG An ,ZHANG Min (School of Computer Science and Communication Engineering, China University of Petroleum, Dongying 257061)   

  • Online:2018-11-16 Published:2018-11-16

摘要: 提出了一种基于约束规划的选择性神经网络集成方法,在训练出个体网络之后,用约束规划方法选择出相对最佳的个体网络组成神经网络集成。理论分析和实验结果表明,该方法设计过程简单,能够以较小的运算代价提高神经网络集成的泛化能力。

关键词: 神经网络 神经网络集成 约束规划

Abstract: A selective constructing approach to neural network ensemble based on constrained programming is presented. Constrained programming is used to select the trained individual networks to construct a new ensemble. Theoretical analyses and experimental result

Key words: Neural network, Neural network ensemble, Constrained programming

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!