用人工智能方法估计东南沿海地震带的地震活动趋势

本文提出了对东南沿海地震带地震活动趋势的一个预测，采用了人工神经元网络技术来获取有关地震活动涨落起伏韵律性的知识。网络训练的样本是用一个沿地震时间历程滑动的时间窗采样获取的。网络经济足够多次数的学习，很好地记忆了资料序列中蕴含的时序特征。结果表明，东南沿海地震带的地震活动将在当前的剩余释放阶段延续二十年左右，然后转入下一个积累释放阶段。

THE ESTIMATION OF SEISMICITY TENDENCY ATTHE SEISMIC ZONE OF SOUTHEASTER COAST BYTHE METHOD OF ARTIFICAL INTELLIGENCE

A seismic tendency at the seismic zone of southeastor coast is forecasted in this paper. For acquiring the knowedge of the seismicity rhythm, the artifical neural network technique is adopted. The samples for training the network are constructed by a sampling window moving through the time history of earthquake. After learning sufficiently in many times, the network can memorize the characteristics of the data scries quite well. The result shows that the seismicity in the zone will continue in current remnant release stage about twenty years or more than, and then it turns into the next accumulating release stage.