混沌RBF神经网络的地磁变化场预测模型

Prediction Model of the Geomagnetic Variation Field by Chaotic RBF Neural Network

提出一种融合混沌理论和RBF(radial basis function)神经网络的地磁变化单站预测模型。分析磁场数据的混沌特性,求取关键参数嵌入维数m和时间延迟τ,据此对初始数据进行相空间重构,并将经混沌理论优化的样本集作为神经网络的训练集和测试集进行仿真实验。结果表明,经混沌理论改进后的RBF神经网络模型可以较为准确地预测地球磁场的变化趋势,对我国地磁场的适用性较好,具有一定的泛化能力。

We propose a single station prediction model of geomagnetic variation based on chaos theory and RBF neural network. We analyze the chaotic characteristics of magnetic field data, and obtain the embedding dimension m and time delay τ.Based on this, we reconstruct the phase space. The sample set optimized by chaos theory is used as the training and test set of the neural network for simulation experiment. The results show that the RBF neural network model improved by chaos theory can accurately predict the change trend of geomagnetic field and has good applicability to geomagnetic field in China.