改进主成分分析方法及其在地震数据处理中的应用
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
本文基于地震数据的特征,引入向量成分化概念,提出了一种改进的PCA方法,并给出了相应的算法实现。应用改进后的PCA方法,分析了西安高陵2009年11月5日地震的前兆数据以及之前发生在西安地区的多次地震的前兆数据。分析结果表明:这种改进的PCA方法在处理地震数据时,主成分数据拟合曲线的震前特征更便于数学表示,为地震的分析和预测提供了良好的基础。
This paper introduces the concept of vecter of proportions based on the characteristics of seismic data and presents an improved method of principal component analysis and a way to carry out it.Then we apply the improved PCA method to the precursor data of Gaoling earthquake which happened in Xi'an in November 5,2009,as well as other precursor datas of earthquakes which happened in Xi'an before.The results show that this improved PCA method makes the pre-earthquake characteristics lie in the principal component data fitting curve more easier to express in mathematic when processing seismic data and that provides a good basis for earthquake analysis and prediction.
引文
[1]张军.时间序列数据中的模式挖掘及其在地震预报中的应用研究.计算机软件与理论,2006,3.
    [2]HOTELLING H.Analysis of a complex of statistical variables intoprincipal components[J].Journal of Educational Psychology,1933,24:417-441.
    [3]Moghaddam B1 Principal manifolds and probabilistic subspaces forvisual recognition[J].1 IEEE Transactions on Pattern Analysis and MachineIntelligence,2002,24,(6):780-788.
    [4]印兴耀,孔国英,张广智.基于核主成分分析的地震属性优化方法及应用.石油地球物理勘探,2008,43,(2):179-183.
    [5]余锦华,杨维权,多元统计分析与应用.中山大学出版社,210-206.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心