基于SOM和BP神经网络的地震预报技术

摘要

选取闽粤赣交界地区及东南沿海作为研究对象,以测震学前兆指标作为预报因子,使用SOM网络对地震预报因子进行分类,分别构建BP网络进行学习和预测,克服评测样本数量有限且分布不均匀给测试带来的困难.结果表明,该方法的预测精度优于单一采用BP网络的精度,对地震预报具有一定应用价值.%The border area among Fujian, Guangdong and Jiangxi provinces and southeastern coast area are chosen as the research object. The indices of seismological precursor are used as predictors. Firstly, SOM neural network is used to classify the predictors and then the samples are studied and predicted by establishing BP neural network respectively, which can overcome the problems of the limitation and unevenly distribution of the number of samples. Simulation results show that the prediction accuracy of this method is better than that of just one of BP network, which is quite valuable to the prediction of earthquake.