基于移动窗口和动态优化的支持向量回归在指数预测中的应用  

APPLYING SUPPORT VECTOR REGRESSION BASED ON MOVING WINDOW AND DYNAMIC OPTIMISATION IN INDEX PREDICTION

作　　者：简清明[1] 曾黄麟[2] 叶晓彤[1] 

机构地区：[1]四川理工学院网管中心,四川自贡643000 [2]四川理工学院人工智能四川省重点实验室,四川自贡643000

出　　处：《计算机应用与软件》2011年第12期83-85,101,共4页Computer Applications and Software

基　　金：国家高技术研究发展计划基金(2008AA11A134)

摘　　要：证券投资是一种高风险、高收益的理财方式,而对市场运行趋势的良好把握有助于投资者回避风险、把握机会。提出基于移动窗口和边界动态缩放的PSO动态优化的在线支持向量机回归方法,预测沪深300指数的短期走势。模拟实验证明,该方法和传统学习方法相比,在测试的精度和实时性上都有较大的提高。Securities investment is a high risk,high-yield financial management method,while the good grasp of market trends helps the investors avoid risk and seize the opportunity.In this paper,we propose a method based on PSO dynamical optimisation of moving window and boundary dynamic zooming as well as online support vector machine recession algorithm to predict short term trend of Shanghai and Shenzhen Stock Index 300.Simulative experiments show that the method significantly improves the test in both precision and real-time compared with traditional learning method.

关 键 词：证券投资 移动平均线 支持向量机 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]