人工神经网络在地震中短期预报中的应用

The Application of Artificial Neural Networks in Mid-and Short-term Earthquake Prediction

本文将BP神经网络用于地震中短期预报。作者把一些常用的地震学指标作为神经网络的输入 ,而将BP神经网络的输出作为表征地震活动平静的特征参数Wq,井将其用于华北地区进行空间扫描 ,结果表明中强地震前 1年左右或稍长时间 ,未来震中周围一般都开始出现Wq 值的中短期异常区 ,证明本方法具有很好的中短期预报效果。

BP neural networks is used for the mid\|and short\|term earthquake prediction in this paper.Some prediction parameters of seismology are the input units of neural networks.And the output units of neural networks is called as the character parameter W q described quiescence of seismicity.It is used to scan in North China.The result shows that the mid\|and short\|term anomaly zone of W q \|value usually appeared obviously around the future epicenter before earthquake about one year or longer with good prediction ability.