上证综指买卖策略算法的设计与实现  

DESIGN AND IMPLEMENTATION OF ALGORITHM OF SCI TRADING STRATEGIES

作　　者：贾国芳[1] 

机构地区：[1]广东技术师范学院计算机与网络中心,广东广州510665

出　　处：《计算机应用与软件》2008年第6期249-251,共3页Computer Applications and Software

摘　　要：将改进的多层BP神经网络预测模型与上证综指买卖策略算法相结合得到一个新算法。模拟9年上证综指的结果表明:由该算法买卖指数,投资者获得的回报比持有策略获得的回报大约高三倍。This paper introduced a new algorithm which combining improved multi-layer back-propagation （BP） neural network forecasting model with Shanghai Composite Index （SCI） Trading Strategies. The result of simulating nine years of SCI transaction shows that investors＇ return earned by using this algorithm is about three times as large as that earned by the buy-and-hold strategy.

关 键 词：BP神经网络 时间序列 SCI 核平滑 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TP311.1[自动化与计算机技术—计算机科学与技术]