摘要
简介了新一代地震预报专家系统NGESEP、BP神经网络模型及其算法。由于BP神经网络具有很强的非线性映射功能，它可以很好地反映震前出现异常的种类和异常时间与未来地震震级之间的较强非线性关系。在NGESEP系统中可以从实例库中提取典型震例并通过BP网络进行学习。
The new generation expert system for earthquake prediction (NGESEP), the modelof BP neural networks and its algorithm are briefly introduced in the paper. It can wellrepresent the nonlinear relation between future earthquake magnitude and anomaly cat-egories and anomaly time because BP neural networks have stronger function ofnonlinear imaging. In NGESEP the typical earthquake cases can be subtracted from thecase database and converted to the sample data to train the BP networks. The examina-tion of the practical examples indicates that the prediction result for the forthcomingearthquake magnitudes is very good.
出处
《地震》
CSCD
北大核心
1997年第2期142-148,共7页
Earthquake
基金
地震科学联合基金!92271
