基于遗传算法的小波神经网络在地震形变前兆分析中的应用  

WAVELET NEURAL NETWORK BASED ON GENETIC ALGORITHM IN THE APPLICATION OF SEISMIC DEFORMATION PRECURSOR ANALYSIS

作　　者：袁国旭[1] 任峻峰[1] 李文军[1] 王向亮[2] 

机构地区：[1]河北省地震局承德地震台,河北承德067000 [2]河北省地震局宽城地震台,河北宽城067600

出　　处：《高原地震》2016年第4期49-54,共6页Plateau Earthquake Research

摘　　要：利用遗传算法来优化小波神经网络,达到提高逼近精度,简化网络结构,并利用优化后的小波神经网络训练地震前兆形变资料,运用MATLAB对拟合和预测过程进行仿真。结果表明,该方法对地震预测有参考作用。The wavelet neural network is optimized by genetic algorithm in order to improve the approximation accuracy and simplify network structure, and the results are used to train the earthquake precursor deformation data. The fitting and forecasting process are simulated by the MATLAB. The results show that the method has reference function for earthquake prediction.

关 键 词：遗传算法 小波 神经网络 地震预测 

分 类 号：P315[天文地球—地震学]