Application and Research on SOM-PCA Based RBF Neural Network in Earthquake Prediction
The problems of sample dispersion and information overlap of the earthquake prediction factors cause the insufficient precision when predict directly via neural network. For this problem, a Radial Basis Function (RBF) neural network prediction model algorithm based on Self-Organizing Map (SOM) neural network and Principal Components Analysis (PCA) is introduced in this article. Firstly, the earthquake prediction factors are divided in several categories by SOM neural network. And then, the dimensionality of the data in each category is descended by PCA respectively. At last, an RBF neural network whose input is the principal components of the data is trained to predict earthquake in each category. The simulation results show that, compare with the traditional prediction method via neural network directly, the proposed scheme can effectively improve the prediction accuracy.
Earthquake Prediction SOM Neural Network Principal Components Analysis RBF Neural Network Earthquake Magnitude
Yi Chen Ying Wang
School of Computer and Control Guilin University of Electronic Technology, Guilin, China
国际会议
哈尔滨
英文
315-320
2010-08-01（万方平台首次上网日期，不代表论文的发表时间）