基于序列匹配和极值理论对地震趋势的研究  

Research on seismic tendency based on sequence matching and extreme value theory

作　　者：邱剑锋[1] 谢娟[2] 李炜[1] 汪继文[1] 周健[1] 

机构地区：[1]安徽大学计算机科学与技术学院,安徽合肥230039 [2]安徽建筑工业学院数理系,安徽合肥230022

出　　处：《计算机工程与设计》2011年第5期1841-1844,共4页Computer Engineering and Design

基　　金：安徽省自然科学基金项目(090412063);安徽省教育厅优秀青年基金项目(2010SQRL018);安徽大学青年科学研究基金项目(2009QN027B)

摘　　要：通过分析造成地震发生的各种因素之间的关系,提出了将极值理论和地震时间序列匹配相结合用于地震趋势研究的基本思想。针对传统地震时间序列匹配算法的不足,提出了将基于形态的时间-震级二维相似性匹配算法和极值理论相结合对未来地震趋势进行分析的方法。该方法采用时间-震级二维相似性匹配的方法寻找最佳地震匹配区域,用极值理论对该区域地震活动的复发周期和发震概率及其概率阈值进行数据处理,从而获得对该地区未来一段时间的地震趋势做出判断。以安徽及周边地区为例,阐述了应用该方法进行研究的一般步骤。Based on the analysis of the relations among various factors contributing to earthquake occurrences, a basic thought is put forward that the combination of extreme value theory （EVT） with a time-sequence similarity matching algorithm for seismological relevant zones is adopted for earthquake trends. To solve the deficiency of traditional time-sequence similarity matching algorithm, a new method is developed combining morphology-based time sequence - earthquake magnitude 2D similarity matching with EVT. Time sequence - earthquake magnitude 2D similarity matching method is used to determine optimum seism matching zones, while EVT is ado- pted for processing data on earthquake recurrence period, occurrence probability and probability threshold in the region to effectively predict earthquake trends for some time to come in the region. The general procedures for application of this method on earthquake study are presented, taking Anhui and surrounding areas for example.

关 键 词：粗粒度 细粒度 时间-震级二维相似性 时间序列匹配 极值理论 

分 类 号：TP339[自动化与计算机技术—计算机系统结构]