基于SVM 的橡胶连续收盘指数的降噪回归预测
PDF下载 (479)邵明星,王保山,丁爱亮,徐炳振*.基于SVM 的橡胶连续收盘指数的降噪回归预测[J].宁波大学学报(理工版),2012,25(03):75-78.DOI:
SHAO Ming-xing,WANG Bao-shan,DING Ai-liang,XU Bing-zhen*.On Consecutive Closing Price of Rubber Based on SVM’s Regression Forecast with Noise-reduction[J].Journal of Ningbo University(Natural Science & Engineering Edition),2012,25(03):75-78.DOI:
| Title: | On Consecutive Closing Price of Rubber Based on SVM’s Regression Forecast with Noise-reduction |
| 作者: | 邵明星, 王保山, 丁爱亮, 徐炳振* |
| Author(s): | SHAO Ming-xing, WANG Bao-shan, DING Ai-liang, XU Bing-zhen* |
| 关键词: | 支持向量机; 降噪; 回归预测; 橡胶连续 |
| Keywords: | support vector machine; noise reduction; regression prediction; rubber continuous |
| 分类号: | O411.1; F035.4 |
| 文献标识码: | A |
| 摘要: | 基于支持向量机的预测模型对橡胶连续的收盘价原始数据和降噪后数据进行预测, 并对 拟合得到的2 个结果进行对比. 结果表明: 对时间序列降噪后, 再用支持向量机进行预测模拟, 相比直接用原始数据得到预测结果要精确得多, 也具有很好的推广价值. |
| Abstract: | Based on support vector machine, the raw data and data with noise reduction are separately used to predict rubber consecutive closing price, and comparison of these two results is conducted and conclusion is obtained. The study suggests that the forecast using the data of time series with noise reduction is more accurate than the one directly with the raw data. |
| 参考文献 /References: | [1] Farmer J D, Sidorowich J. Predicting chaotic time series
[J]. Phys Rev Lett, 1987, 59:845-848.
[2] 甘建超, 肖先赐. 基于相空间邻域的混沌时间序列自
适应预测滤波器(II)非线性自适应滤波[J]. 物理学报,
2003, 52(5):2995-2999.
[3] Vapnik V N.统计学习理论的本质[M]. 张学工, 译. 北
京: 清华大学出版社, 2000.
[4] 曹丽娟, 王小明. 金融工程的支持向量机方法[M]. 上
海: 上海财经大学出版社, 2007.
[5] Cortes C, Vapnik V. Support vector network[J]. Machine
Learning, 1995, 20: 273-297.
[6] 徐建华, 张学工, 李衍达. 支持向量机的新方法[J]. 控
制与决策, 2004, 19(5):481-485.
[7] 冯振华, 杨杰明. SVM 回归参数探讨[J]. 机械工程与
自动化, 2007(3):17-22.
[8] Cao L, Hong Y, Fang H, et al. Predicting chaotic time
series using wavelet network[J]. Physica D, 1995, 85(1):
225-238.
[9] Atsalakis G, Valavanis K. Surveying stock market forecasting
techniques, Part II: Soft computing methods[J].
Expert Systems with Application. 2009, 36:5932- 5941. |
| 备注/Memo: | 收稿日期: 2012?02?20. 宁波大学学报(理工版)网址: http://3xb.nbu.edu.cn 基金项目: 国家自然科学基金(10774080). 第一作者: 邵明星(1986-), 女, 山东临沂人, 在读硕士研究生, 主要研究方向: 金融物理与复杂系统动力学. E-mail: shmx1213@126.com *通讯作者: 徐炳振(1959-), 男, 山东德州人, 博士/教授, 主要研究方向: 金融物理与复杂系统动力学. E-mail: xubingzhen@nbu.edu.cn 宁波大学学报(理工版)网址:http://journallg.nbu.edu.cn/ |