第4期 钱晓山,等:改进基因表达式编程在股票中的研究与应用 307· 800r 700 model based on a neural network-expert system Hybrid 600 [C]//Proeedings of the IEEE International Conference on 500 400 真实值 Neural Network.Seattle,1991::1289-1293. 300 ……预测值 200 [8]PESARAN M H,TIMMERMANN A.A recursive modeling i00 0 approach to predicting UK stock C]//Economic Journal. 5 1015 202530 时间天 Blackwell Publishing.Edward 2000::159-191. 9)]元昌安,唐常杰,左劫,等.基于基因表达式编程的函数 图4IGEP算法上证指数部分训练数据真实值与模型 挖掘一收敛性分析与残差制导进化算法[J刀.四川大学 值比较曲线图(2004年前30天) 学报:.工程科学版,2004,36(6):100-105. Fig.4 The IGEP algorithm for the compared curve between YUAN Changan,TANG Changjie,ZUO Jie,et al.Func- the true value and model values of the Shanghai in- tion mining based on gene expression programming-conver- dex part training data(formerly 30 days in 2004) gence analysis and remnant-guided evolution algorithm[J]. 4 结束语 Journal of Sichuan University:Engineering Science Edition, 2004,36(6):100-105 用GEP进行预测,不需要知道各因素间的因果 [0]段磊,唐常杰,左哉,等.基于基因表达式编程的抗噪声 关系,只需要提供足够的实验或者实验数据,无须知 数据的函数挖掘方法[刀,计算机研究与发展,2004,41 道目标函数,就可以达到准确预测的目的.本文 (10a1684-1689.. 创新点在于提出了一种新的IGEP算法,可得到准 DUAN Lei,TANG Changjie,Zuo Jie,et al.An anti-noise 确的函数表达式,通过实例计算和分析可知,此方法 method for function mining based on CEP[J]].Journal of 优于文献[11]的GEP方法,并充分发挥了进化算法 Computer Research and Development,2004,41(10)): 1684-1689., 内含的并行性,从而使得算法在求解速度上远远快 [11]廖勇.基于基因表达式编程的股票指数和价格序列分 于普通的算法 析[D.成都::四川大学,2005 参考文献: LIAO Yong,Time series prediction in stock-price index and stock-price based on Gene Expression programming [1]FERREIA C.Gene expression programming:a new adaptive [D].chengdu:Sichuan University,2005.. algorithm for solving problems [J].Complex Systems, [12]李曲,蔡之华,朱利等.基因表达式编程方法在采煤工 2001,13(2)87-129. 作面瓦斯涌出量预测中的应用[刀]应用基础与工程科 [2]MITCHELL M.An introduction to genetic algorithms[M]. 学学报,2003,12(1))505 Cambridge::IMIT Press,1996::143-164. LI Qu,CAI Zhihua,ZHU Li,et al.Application of gene 3 ]FERREIRA C.Gene expression programming:mathematical expression programming in predicting the amount of gas e- modeling by an arifcial intelgence[M].2nd Ed.Mit mited from coal facefJ].Journal of Basic Science and Press Cambridge,Massachusetts o London,England Spring-g- Engneering,2003,12(1)):05 er2006:f82-85.5, 作者简介: [4]张克俊.求解反问题的改进的基因表达式编程研究[D] 钱晓山,男,1980年生,讲师,博士 南鲳:江西理工大学,2006.6. 研究生,主要研究方向:复杂工业过程 ZHANG Kejun.The study of improving gene expression pro-o- 建模、优化控制 gramming for solving the inverse problem[D]].Nanchang: liangxi University of science and technology,2006.6 [5]ZUO Jie,TANG Changjie,LI Chuan,et al.Time series prediction based on gene expression programming[Cl//In-n- termational Conference for Web Information Age 2004.Ber-er 阳春华,女,1965年生,教授,博士 ling Heidelberg:Springer Verlag 2004:5564.4. 生导师,获”湖南省青年科技奖”、”湖 6 ]]ECONOMIC W H.Prediction using neural networks:the 南省十大杰出女性”,主要研究方向为 case of IBM daily stok returns[C]//IEE Internationalal 复杂工业过程建模、优化控制:智能信 Conference on Neural Network.1988:3451458 58. 息处理 [7]]BERGERSON K,WUNSCH D C.A commodity trading:[8] PESARAN M H,TIMMERMANN A. A recursive modeling approach to predicting UK stock [ C] //Economic Journal, Blackwell Publishing,Edward 2000: 159-191. ZHANG Kejun. The study of improving gene expression programming for solving the inverse problem[ D] . Nanchang: liangxi University of science and technology,2006. 参考文献: 500 5 [2] MITCHELL M. An introduction to genetic algorithms[ M] . Cambridge: MIT Press,1996: 143-164. 作者简介: 真实值 ………预测值 800 600 LIAO Yong,Time series prediction in stock-price index and stock-price based on Gene Expression programming [D] . chengdu: Sichuan University,2005. 400 200 用 GEP进行预测,不需要知道各因素间的因果 关系,只需要提供足够的实验或者实验数据,无须知 道目标函数,就可以达到准确预测的目的.本文 创新点在于提出了一种新的IGEP算法,可得到准 确的函数表达式,通过实例计算和分析可知,此方法 优于文献[11] 的 GEP方法,并充分发挥了进化算法 内含的并行性,从而使得算法在求解速度上远远快 于普通的算法. [5] ZUO Jie,TANG Changjie,LI Chuan,et al. Time series prediction based on gene expression programming[ C]//Intermational Conference for Web Information Age 2004. Berling Heidelberg: Springer Verlag,2004: 5564. 钱晓山,男,1980年生,讲师,博士 研究生,主要研究方向: 复杂工业过程 建模、优化控制. LI Qu,CAI Zhihua,ZHU Li,et al. Application of gene expression programming in predicting the amount of gas emited from coal face[J] . Journal of Basic Science and Engneering,2003,12(1) :50-5. [4] 张克俊.求解反问题的改进的基因表达式编程研究[D] . 南昌: 江西理工大学,2006. O 700 [7] BERGERSON K,WUNSCH D C. A commodity trading [11]廖勇.基于基因表达式编程的股票指数和价格序列分 析[D] .成都: 四川大学,2005. Fig.4 The IGEP algorithm for the compared curve between the true value and model values of the Shanghai index part training data(formerly 30 days in 2004) 3 ] FERREIRA C. Gene expression programming: mathematical modeling by an arifcial intelgence[ M].2nd Ed. Mit Press Cambridge,Massachusetts o London,England Springer 2006: 82-85, 6 ] ECONOMIC W H. Prediction using neural networks: the case of IBM daily stok returns[ C] //IEE International Conference on Neural Network. 1988: 451458 12] 李曲,蔡之华,朱利等.基因表达式编程方法在采煤工 作面瓦斯涌出量预测中的应用[J]].应用基础与工程科 学学报,2003,12(1) :50-5 9] 元昌安,唐常杰,左劫,等.基于基因表达式编程的函数 挖掘一收敛性分析与残差制导进化算法[J] .四川大学 学报:工程科学版,2004,36(6) :100-105. [1] FERREIA C. Gene expression programming: a new adaptive algorithm for solving problems [J]. Complex Systems, 2001,13(2) : 87-129 图4 IGEP算法上证指数部分训练数据真实值与模型 值比较曲线图(2004年前30天) DUAN Lei,TANG Changjie,Zuo Jie,et al. An anti-noise method for function mining based on CEP[J] . Journal of Computer Research and Development,2004,41(10) : 1684-1689. 307· 4 结束语 model based on a neural network-expert system Hybrid [C]//Proeedings of the IEEE International Conference on Neural Network,Seattle,1991: 1289-1293. YUAN Changan,TANG Changjie,ZUO Jie,et al. Function mining based on gene expression programming-convergence analysis and remnant-guided evolution algorithm[J] . Journal of Sichuan University: Engineering Science Edition, 2004,36(6) :100-105. 第4期 1 0 15 2 0 2 5 3 0 i00 [10] 阳春华,女,1965年生,教授,博士 生导师,获"湖南省青年科技奖"、"湖 南省十大杰出女性",主要研究方向为 复杂工业过程建模、优化控制;智能信 息处理. 段磊,唐常杰,左哉,等.基于基因表达式编程的抗噪声 数据的函数挖掘方法[J] .计算机研究与发展,2004,41 (10) :1684-1689. 时间/天 钱晓山,等: 改进基因表达式编程在股票中的研究与应用 300