摘要
支持向量机回归理论与神经网络等非线性回归理论相比具有许多独特的优点。SVMR有线性回归和非线性回归,其模型的选择包括核的选择、容量控制以及损失函数的选择。SVMR在控制方面的研究包括非线性时间序列的预测及应用、系统辨识以及优化控制和学习控制等方面的研究。将SVMR应用于控制方法的研究,是智能控制的一个崭新的研究方向,具有广阔的应用前景。
The support vector machines regression shows excellent performanc by comparison with other non-linear regression, such as neural networks. SVMR includedes linear SVMR and nonlinear SVMR. The way of SVMR model selection includes kernel selection, capacity selection and loss function selection. The research of SVMR control is currently perdiction of nonlinear time series and its application, system identification, optimal control and learning control. The research of control method using SVMR is a new research field of intelligent control, and will widely be applied in engineering.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2003年第2期192-197,共6页
Pattern Recognition and Artificial Intelligence