摘要
针对卫星地面站资源配置问题输入输出关系的复杂性,提出支持向量机（SVM）回归模型。在模型学习过程中,利用正交设计抽样抽取小数量样本对SVM进行训练,求解二次规划问题,根据测试结果对模型参数进行调整以实现较小的预测误差。实验结果表明,该方法具有较好的拟合精度和泛化能力。
An SVM regression model is proposed to deal with input/output relation complexity of satellite ground station resources allocation.During the process of model learning,small size samples are used by orthogonal design sampling to train SVM to solve quadratic programming problem.The model parameters are adjusted based on test results to keep prediction error small.Experiment results show that the proposed method has good fitting accuracy and generalization ability.
出处
《飞行器测控学报》
2011年第2期15-19,共5页
Journal of Spacecraft TT&C Technology
关键词
资源配置
支持向量机回归
二次规划
拟合
Resources Allocation
SVM Regression
Quadratic Programming
Fitting
