主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 1997, Vol. ›› Issue (3): 32-36.

• 论文 • 上一篇    下一篇

基于遗传技术辅助设计的神经网络期货市场预测

冯旭东, 陈方   

  1. 中国科技大学管理科学系, 合肥, 230027
  • 收稿日期:1997-05-30 出版日期:1997-09-28 发布日期:2012-03-06

Futures Trading Market Prediction Using Neural Network Based on Genetic Algorithms Learning

Feng Xudong, Chen Fang   

  1. University of Science and T6chnology of China, Hefei
  • Received:1997-05-30 Online:1997-09-28 Published:2012-03-06

摘要: 本文在对期货市场的历史数据进行预分析的基础上,建立了神经网络期货市场预测模型.文中不仅对神经网络进行了改进研究,还利用遗传技术优化网络的结构和参数.运运实例对模型进行学习与测试的实验结果表明,利用遗传技术辅助设计的神经网络预测模型能较准确地预报期货价格的波动趋势.

关键词: 期货市场预测, 神经网络, 遗传算法, 辅助设计, 时间序列模型

Abstract: This paper discusses the problems about prediction based on poeural networks firstly,then proPOses a modified training algorithms. Furthermore, it describes the method and procedure of applying genetic algorithms to designing a multi-layer feedforward neural network. Finally, a neural network model of futures trading market prediction has been developed. The results of experiment have shown the effectiveness of this method.

Key words: Futures trading market prediction, Neural networks, Genetic algorithrns, Designing, Time series models