【摘要】 针对优化径向基函数神经网络的各参数问题,提出一种动态K均值混合优化RBF神经网络并应用于广西降水数据进行建立预报模型,该模型与传统的K均值RBF模型和同期的T213降水预报进行对比,结果表明,该模型建立的5月3个区域的逐日降水预报,预测的精确度明显高于同期的T213降水预报。更多还原

【Abstract】 Aiming at the problem of parameter optimization of RBF neural network, presents a dynamical K-means algorithm hybrid optimization of RBF neural network and applies precipitation forecast model in Guangxi. The result shows that prediction accuracy of the may three regional RBF neural network models are significant higher than T213 in the corresponding period for daily rainfall forecast.更多还原