摘要
本文简述了自组织竞争模型的原理,研究了其在地震预报样本分类中的应用。并采用遗传算法对其网络结构参数进行优化,对输入样本参数进行优选。震例检验结果显示自组织竞争网络具有较好的分类效果,经遗传算法优化后的模型分类效果得到进一步改善。
In this paper, the principle of self-organization competition artificial neural network (SCANN) is summarized and its application to earthquake sample classification is studied. The genetic algorithm (GA) is also used to optimize network structure and select input parameters. The results verified by an earthquake example show that SCANN has preferable classification effect and that of SCANN optimized with GA has a further improvement.
出处
《东北地震研究》
2003年第4期18-23,共6页
Seismological Research of Northeast China