RBF（径向基神经网络）网络是一种重要的神经网络，RBF网络的训练分为两步，第一步是通过聚类算法得到初始的权值，第二步是根据训练数据训练网络的权值。RBF权值的初始聚类方法较为复杂，比较简单的有K均值聚类，复杂的有遗传聚类，蚁群聚类等，这个RBF网络的程序是基于K均值聚类的RBF代码。(RBF (radial basis function network) is an important neural network. The training of RBF network is divided into two steps. The first step is to get the initial weight by clustering algorithm, and the second step is to train the weight of the network according to training data. The initial clustering method of RBF weights is relatively complex. There are relatively simple K mean clustering, complex genetic clustering, ant colony clustering, etc. this RBF network is based on K means clustering RBF code.)