讨论了基于径向基函数 ( RBF)的概率神经网络的基本网络结构和网络的学习和运行过程 ,并且与 BP算法的径向基神经网络进行了对比 ,同时也测试了网络的容错能力 .结果表明 ,基于RBF的概率神经网络 ,学习速度大大提高 ,同时减小了 BP陷入局部极小的问题 ,有一定的抗噪声能力 .基于 RBF的概率神经网络模型在心律失常自动识别中获得了很好的应用 The basic network structure and network learning and operation process of probabilistic neural network based on Radial Basis Function (RBF) are discussed and compared with RBF neural network, and the network fault tolerance is also tested.Results The results show that the RBF-based probabilistic neural network can greatly improve the learning speed, reduce the problem of BP falling into a local minimum, and have certain anti-noise ability. The RBF-based probabilistic neural network model is very good in the automatic identification of arrhythmia Applications