【作者】 程砚秋； 杨德权；

【Author】 CHENG Yan-qiu YANG De-quan (School of Management, Dalian University of Technology, Dalian 116023, China)

【机构】 大连理工大学管理学院；

【摘要】 提出一种将决策树和支持向量机相结合的金融预测方法。此方法首先利用决策树筛选出重要属性,然后用线性核和高斯核支持向量机分析了金融时间序列的特点,并构造了适当的核函数,最后根据选择的输入变量和构造的核函数作出最终的预测。通过对深证个股日收盘价和人民币汇率月均价的实证研究,发现此方法具有很好的预测效果。更多还原

【Abstract】 This paper presents an approach based on decision tree (DT) and support vector regression (SVR). The method firstly selects the significant attributes based on decision tree, and then analyses the financial time series through SVR with linear kernel function or Gauss kernel function, and then constructs a fusion kernel function, and finally predicts the result. This study applies DT-SVR to predict stock prices coming from Shenzhen stock market and RMB exchange rates. Experiments show that, t DT-SVR has a very good prediction accuracy.更多还原