使用人工神经网络进行我国大陆强震时间序列预测  

FORECASTING METHOD OF EARTHQUAKE TIME SERIAL IN MAINLAND OF CHINA BY USING THE ARTIFICIAL NEURAL NETWORKS

作　　者：王炜[1] 谢端[2] 宋先月[1] 赵利飞[1] 

机构地区：[1]上海市地震局,上海200062 [2]中国科学技术大学地球与空间科学系,安徽合肥230026

出　　处：《西北地震学报》2002年第4期315-319,共5页Northwestern Seismological Journal

基　　金：地震科学联合基金(100028)

摘　　要：使用人工神经网络对中国大陆最大地震时间序列进行预测,预测次年的中国大陆最大地震震级,结果表明该方法具有较好的预报效果.还表明中国大陆强震活动除了与强震时间序列本身有关外,还与全球的强震活动、太阳黑子活动等有密切的非线性关系.The artificial neural networks is used to prediction of time serial of the earthquake max.magnitudes in mainland of China,and to forecast earthquake max.magnitude for next year.The result shows that the method has good forecasting ability.The result also indicates that seismicity of strong earthquake in mainland of china is not only related with the time serial of strong earthquake but also the seismicity of strong earthquake in the world and the sunspot.

关 键 词：地震时间序列 人工神经网络 太阳黑子 地震预测 震级 

分 类 号：P315.75[天文地球—地震学]