给出了一种基于最小二乘支持向量机(LS-SVM)的码分多(CDMA)系统信号接收机非线性自适应核算法。LS-SVM使用均方误差原则,优点是能应用于自适应在线的情况,基于LS-SVM的接收检测器复杂度适中,能够实现非线性分类,并能够自适应执行,而且仅仅只需要从用户那里得到较少的训练数据序列。
This paper discusses an adaptive kernel nonlinear learning algorithm which is based on LS- SVM for Code Division Multiple Access (CDMA) system. LS- SVM uses mean squared error criterion. An advantage of it is that the algorithm can be implemented adaptively on - line. The LS - SVM detectors have advantages in that they have moderate complexity, can be implemented adaptively, and require only training sequence data from the desired user.