火力与指挥控制2013，Vol.38Issue(3)：130-132,140,4.
ε-支持向量回归的噪声敏感性研究
On the Noise Sensitivity of ε-support Vector Regression
摘要
Abstract
Aiming at the question of ε -support vector regression (ε -SVR) for noise date sensitirity, the Bayesian estimation theory is used to discussed the relationship between data noise level and insensitive parameter ε. The optimization problem of ε-SVR is interpreted into Bayesian estimation problem, to discuss the noise sensitivity of ε-SVR. Simulation results show different intensity of noise makes greater impact on the effect of the return,selecting the appropriate parameter e can reduce the sensitive to noise, meanwhile proves the theoretical derivation of effectiveness of Bayesian estimation theory.关键词
ε-支持向量回归/贝叶斯估计/不敏感参数/敏感性Key words
ε-support vector regression, Bayesian estimation, insensitive parameter, sensitivity分类
信息技术与安全科学引用本文复制引用
田韶超,黄景涛,李广义..ε-支持向量回归的噪声敏感性研究[J].火力与指挥控制,2013,38(3):130-132,140,4.
