支持向量机在射频功率器件建模中的应用

The application of support vector machine in RF power device modeling

针对现有射频功率器件建模方法的不足,运用支持向量机对射频功率器件进行建模.通过软件仿真对比了支持向量机和神经网络的不同结果,得出支持向量机建立的模型精确度更高,更适合小样本条件下的建模.并且针对实际测试中出现的特殊情况,提出引入领域知识的方法,将散射函数具有的约束条件同支持向量机结合,使得支持向量机具有相关领域知识的支撑.比较了原始和领域知识支持向量机建模的不同结果,得出领域知识支持向量机在该种情况下具有更好的模型精度.

The existing methods for the modeling of（radio frequency,RF）power device are not satisfactory,so the support vector machine is used to build the modeling.The result is that the support vector machine is better than artificial neural networks in the modeling and more suitable for the training which is lack of samples.Considering some special situations in the actual test,domain knowledge is introduced into the support vector machine.The constraint of scattering function is used here,so the support vector machine can get the domain knowledge and have its power.This experiment compared the modeling results of the original and domain knowledge support vector machines.Simulation results show that the domain knowledge support vector machine has a better accuracy.