经网络提出了一种基于蚁群聚类算法的径向基神经网络. 利用蚁群算法的并行寻优特征和挥发系数方法的自适应更改信息量的能力,并以球面聚类的方式确定了径向基神经网络中基函数的位置, 同时通过比较隐层神经元的相似性、合并相似性较为接近的2 个神经元来约简隐含层的神经元,以达到简化径向-via the Internet presents a clustering algorithm based on the ant colony RBFNN. Ant use the parallel algorithm optimization Volatile coefficient characteristics and the method of adaptive changes in the volume of information, spherical cluster and the method of RBF neural network-based functions of the position, By comparing the same time hidden layer neurons of similarity, Merger similarity is closer to the two neurons to Jane about the hidden layer neurons to achieve simplification Radial