东北大学学报（自然科学版）

东北大学学报:自然科学版 ›› 2016, Vol. 37 ›› Issue (7): 942-946.DOI: 10.12068/j.issn.1005-3026.2016.07.007

• 信息与控制 • 上一篇    下一篇

基于距离和时序的层次粒度支持向量回归机

王珏1,2， 乔建忠1， 林树宽1   

  1. (1.东北大学 信息科学与工程学院， 辽宁 沈阳110819; 2.沈阳农业大学 信息与电气工程学院， 辽宁 沈阳110866)
  • 收稿日期:2015-03-10 修回日期:2015-03-10 出版日期:2016-07-15 发布日期:2016-07-13
  • 通讯作者: 王珏
  • 作者简介:王珏(1980-)，女，辽宁沈阳人，沈阳农业大学讲师，东北大学博士研究生; 乔建忠(1964-)，男，辽宁兴城人，东北大学教授，博士生导师； 林树宽(1966-)，女，吉林长春人，东北大学教授.
  • 基金资助:
    国家自然科学基金资助项目(61272177).

Hierarchical Granular Support Vector Regression Based on Distance and Temporal

WANG Jue1,2， QIAO Jian-zhong1， LIN Shu-kuan1   

  1. 1.School of Information Science & Engineering， Northeastern University， Shenyang 110819， China; 2.College of Information and Electric Engineering，Shenyang Agricultural University，Shenyang 110866，China.
  • Received:2015-03-10 Revised:2015-03-10 Online:2016-07-15 Published:2016-07-13
  • Contact: QIAO Jian-zhong
  • About author:-
  • Supported by:
    -

RichHTML

0

PDF (PC)

754

摘要： 针对目前粒度支持向量回归机的粒划算法只考虑了距离因素，引入时序因素，提出适用于金融时间序列的基于距离和时序的层次粒度支持向量回归机(DTHGSVR).该方法首先将训练样本通过核函数映射到高维空间，并在该特征空间中进行初始粒划.然后，通过衡量样本粒与当前回归超平面的距离以及当前样本粒时序的综合因素，找到含有较多回归信息的粒，并通过计算其半径、密度及时序信息进行深层次的动态粒划.如此循环迭代，直到没有粒需要进行深层划分为止.最后，对不同层次的粒进行回归训练.采用提出的基于距离和时序因素的层次粒度支持向量回归机对基金净值进行预测，实验结果表明回归的泛化性有所提高.

关键词: 粒度支持向量回归, 时序, 金融时间序列, 预测, 泛化性

Abstract: Only distance factor is considered in the granular algorithm of granular support vector regression. Temporal factor was introduced simultaneously in granular algorithm. Hierarchical granular support vector regression based on distance and temporal factors(DTHGSVR) was proposed which is applicable for financial time series. The training samples were mapped into the high-dimensional space by mercer kernel， and the samples were divided into some granules initially. Then， the granules which have more regression information was found by measuring the distances between the granules and regression hyperplane and the granule’s temporal factor. By computing the radius， density of granules and the temporal factor， the deeper hierarchical granulation process was executed until no granules was needed to be granulated. Finally， those granules in different granulation levels were trained by SVR. Fund net was forecast by the hierarchical granular support vector regression based on distance and temporal factors. Experimental results showed the generalization performance of regression had been improved.

Key words: granular support vector regression, temporal;financial time series;forecasting;generalization

中图分类号: