针对RBF神经网络学习算法不能很好地确定其隐含层节点数、隐含层节点中心及其半径的问题,利用AP聚类算法无需事先确定聚类数的特点,提出了一种基于AP聚类的RBF神经网络算法。将该算法应用于120t转炉Q235B钢种冶炼过程的终点碳含量和温度预报,预测结果与实际结果比较,表明该算法具有预测精度高等优点,可为类似应用提供借鉴。 RBF neural network learning algorithm can not well determine the number of its hidden layer nodes, the hidden layer node center and its radius. By using the AP clustering algorithm without prior determining the number of clusters, this paper proposes a clustering algorithm based on AP Class RBF neural network algorithm. The algorithm is applied to predict the carbon content and temperature at the end of Q235B steelmaking process of 120t converter. The comparison between the prediction result and the actual result shows that the algorithm has the advantages of high prediction accuracy and can provide references for similar applications.