摘要
基于支持向量回归(Support Vector Regression,简称SVR)的非线性时间序列预测是智能预测的重要前沿课题,在许多领域有着非常广泛的应用前景。文章介绍了SVR基本理论和方法,从金融、电力、交通、旅游等领域的典型应用对基于SVR的非线性时间序列预测进行了综述,分析了目前SVR在核函数、自由参数选择和输入数据处理方面存在的问题及其在应用领域进一步研究的方向。
Support vector regression（SVR）-based nonlinear time series forecasting is an important advanced topic in intelligent forecasting,and has extensive application prospect in many fields.In this paper,the basic theory and method of SVR are introduced.Then a survey is conducted on SVR-based nonlinear forecasting in view of its application in the fields of finance,power,transportation and tourism.Finally,the existing problems of kernel function,free parameters selection and input data handling are analyzed.Meanwhile,the further application fields in the future research are also prospected.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第3期369-374,共6页
Journal of Hefei University of Technology：Natural Science
基金
国家自然科学基金资助项目(71271072
71131002)
高等学校博士学科点专项科研基金资助项目(20110111110006)
安徽高校省级自然科学研究资助项目(KJ2012B097)
国家级大学生创新训练资助项目(201211305013)
关键词
支持向量机
支持向量回归
非线性
时间序列预测
support vector machine
support vector regression（SVR）
nonlinearity
time series forecasting
