基于自适应投影原理 ,本文提出了构造和训练径向基函数网络的自适应投影学习算法 ,该算法能迭代地确定径向基函数的个数、中心的位置以及网络的权系数 ,从而提供了生成径向基函数网络的一个简单和有效的方法 .通过应用于信道均衡器和时间序列预报信号处理领域中的两个例子 ,进一步证明了该算法是行之有效的 . Based on the principle of adaptive projection, this paper proposes an adaptive projection learning algorithm for constructing and training radial basis function networks, which iteratively determines the number of radial basis functions, the center position and the weight coefficients of the network, thus providing A simple and effective method to generate radial basis function networks is further proved by applying two examples in the field of channel equalizer and time series prediction signal processing.