一种基于Multi-SVM的华东地区地震前兆模式识别及预测方法的研究与探索  

RESEARCH AND EXPLORATION ON A METHOD OF EARTHQUAKE PRECURSOR PATTERN RECOGNITION AND PREDICTION IN EAST CHINA BASED ON MULTI SVM

作　　者：陈锡晶 陈德兴 陈丽丽 冯然 邱剑锋[2] CHEN Xijing;CHEN Dexing;CHEN Lili;FENG Ran;QIU Jianfeng(Chuzhou Seismological Bureau,Chuzhou 239000,China;Anhui University,Hefei 230000,China)

机构地区：[1]滁州市地震局,安徽滁州239000 [2]安徽大学,安徽合肥230000

出　　处：《高原地震》2023年第1期54-59,共6页Plateau Earthquake Research

基　　金：安徽省地震局科研合同制项目(项目编号:202201)资助。

摘　　要：以华东地区1970年以来的地震目录为研究对象,基于Multi-SVM技术,对原始地震目录的样本空间进行预处理。通过分组,在每个分组上计算测震学指标,将样本空间映射至特征空间,采用多分类SVM技术对其进行相关的前兆模式挖掘,并在测试集上进行了回溯性检验。实验结果表明,当一段时间内的地震频次较为稳定时,前兆模式挖掘的结果将更为准确。Taking the earthquake catalog in East China since 1970 as the research object,the sample space of the original earthquake catalog is preprocessed based on Multi SVM technology.Through grouping,the seismo⁃logical indicators are calculated on each grouping,the sample space is mapped to the feature space,and the multi classification SVM technology is used to mine the relevant precursor patterns,and the backtracking test is conducted on the test set.The experimental results show that when the earthquake frequency is relatively sta⁃ble for a period of time,the results of precursor pattern mining will be more accurate.

关 键 词：SVM技术 华东地区 地震前兆 

分 类 号：P315.99[天文地球—地震学]