Abstract: Red tide is an anomalous phenomenon and is characterized by abruptness and nonlinearity, so the red tide prediction has been a hotspot in the oceanographic studies. The fundamentals of RBF neural network are briefly introduced, and the application of artificial neural network method to the red tide prediction is discussed. Based on the RBF neural network, the simulation experiments are also illustrated by using red tide monitoring data, and the experimental analysis is also proposed.

Key words: red tide, Radial Basis Function（RBF） neural network, environment factors, red tide prediction

摘要： 赤潮是一种由多因素综合作用引发的生态异常现象，具有突发性及非线性等特点。对其进行预测预报一直是海洋科学研究的热点。简要介绍了RBF神经网络的基本原理，探讨了应用该人工神经网络进行赤潮预测的方法。利用RBF神经网络模型对赤潮灾害监测数据进行仿真实验，并对结果进行了分析。

关键词: 赤潮, 径向基函数（RBF）神经网络, 环境因子, 赤潮预报