摘要
RBF神经网络是前馈神经网络研究中的一个热点.对 RBF神经网络的网络结构、基本原理和学习算法进行了介绍.针对BP神经网络自身的缺陷,提出以RBF神经网络为识别模型,采用最近邻聚类学习算法,建立一种冠心病模式识别诊断系统.仿真实验表明,该模型可快速完成对冠心病样本的学习与拟合,具有预测识别率高的优点,可作为该病诊断的一种有效的辅助手段.
RBFNN has become one of the focuses on feedforward neural networks. This paper introduces the structure, principle and training algorithm of RBFNN. As for the defects of the BP neural networks, a nearest neighbor-clustering algorithm for RBFNN model is presented and applied to the pattern recognition and diagnosis system of coronary heart disease. Simulation experimental results show that the model has good effects on speeding up the learning and approaching process for training sample, with great accuracy. The method can be an efficient aid to diagnosis.
