基于SVM的地震目录最大余震预测算法的设计与实现论文.doc

基于 SVM 的地震目录最大余震预测算法设计与实现

摘 要

随着信息技术的快速发展，地震数据库中的数据量急剧增加。地震数据的海量涌现，给人们对于地

震数据的分析研究带来了极大便利。然而，人们真正需要的是通过对这些已知数据的分析研究，提取真

正直观有效的信息。数据挖掘是数据分析的重中之重，本设计就是根据需要把数据挖掘技术引入到地震

研究领域，讨论数据挖掘的一个核心技术——分类预测，运用一种有效的预测方法——基于 SVM（支持

向量机）的预测回归，研究主震与最大余震之间的关系，预测最大余震，为地震预报提供更有价值的信

息。

首先，对地震数据进行预处理。将哈佛 CMT 地震目录运用最小二乘法解矛盾方程构造拟合函数的方

Abstract

With the rapid development of information technology, the amount of seismic data in the database has

increased dramatically. The massive emergence of seismic data, to analysis and research of seismic data for

people brought great convenience. People really need is, however, through the analysis and study of the known

data, extract the real visual information effectively. Data mining is the key of the data analysis, this design is

based on the need to introduce data mining technology in the field of earthquake research, discuss the a core

technology of data mining, classification prediction, the use of an effective prediction method based on SVM

(support vector machine (SVM) classification prediction, study the main shock and the relationship between

the maximum aftershock, predict the largest aftershocks, provide more valuable information for earthquake

prediction.

Finally, the structure are analyzed, verifying the result of the forecast.

Keywords: Earthquake Catalogue;Support Vector

Machine(SVM);Regression Prediction

目 录

引言........................................................................................................................................................................1

1 绪论....................................................................................................................................................................1

1.1 地震预报....................................................................................................................................................1

1.2 地震目录....................................................................................................................................................3

2 基础知识............................................................................................................................................................3

2.1 机器学习....................................................................................................................................................3

2.2 支持向量机................................................................................................................................................5

4.2

-不敏感函数........................................................................................................................................15

4.3 最优回归超平面与 SVM 线性回归..........................................................................................................16

4.4 非线性 SVM 回归......................................................................................................................................18

4.5 SVM 回归方法的特点...............................................................................................................................18

5 实验模拟和结果评价......................................................................................................................................18

6 结论..................................................................................................................................................................18

致谢......................................................................................................................................................................18

参考文献..............................................................................................................................................................18

1 绪论