摘要
改进了多层BP神经网络预测模型,给出智能交易策略算法,并设计一套智能交易系统以指导投资者买卖上证综合指数。模拟9年上证综指的实验结果表明,由本交易系统获得的回报比持有策略获得的回报大约高三倍。
This paper presents a modified multi -layer back -propagation （BP） neural network forecasting model, and suggests the intelligent trading arithmetic, and designs an intelligent trading system for investors buying and selling Shanghai Composite Index （SCI） The simulation result of nine years to SCI shows that the return achieved by this trading system is about three times as large as that achieved by the buy and hold strategy.
出处
《微计算机应用》
2007年第11期1154-1159,共6页
Microcomputer Applications
关键词
BP神经网络
时间序列
核平滑
BP neural network,time series, kernel smoothing
