用小波变换和最小二乘支持向量机方法预测华南地震区年度最大地震
|
摘要
应用小波变换和最小二乘支持向量机(LSSVM)相结合的方法(小波LSSVM)预测华南地震区年度最大地震。先用小波变换将地震序列分解成不同尺度水平(频率段)的子序列,再用LSSVM方法分别对各子序列建模预测,最后重构各子序列的预测结果并得到最终预测结果。经与周期图方法和LSSVM预测方法比较研究表明:模型输入量中包含地球自转速率变化的小波LSSVM方法预测效果很好,可以用于华南地区年度最大地震预测研究,且地球自转变化与华南地震时间序列的低频部分(趋势)和高频部分(短期变化)之间存在很强的、互不相同的非线性关系。
The models based on wavelet transform(WT) and least squares support vector machine(LS-SVM) was applied to forecast annual the maximum earthquake magnitude in the South China seismic area. Firstly, using wavelet transform the series of annual the maximum earthquake magnitude in the South China seismic area was decomposed into several sub-sequences on different scales(frequency band). Secondly, the least squares support vector machine was applied to forecast the decomposed sub-sequences, respectively. Finally, the reconstruction of forecasted sub-sequences was used as the final predicted result. The comparison with LS-SVM and period method was made. The forecasted results show that WT and LS-SVM method possesses higher prediction precision and excellent forecasting effect, and prove that the method to combine WT with LS-SVM is feasible to forecast time series of strong earthquakes in the South China seismic area. The forecasted results also suggest that there are the strong and different nonlinear relations between variations of velocity of the Earth's rotation with high frequency portion (trend) and low frequency portion(short variations) of time sequence of the annual maximum earthquake in the South China seismic area.
The models based on wavelet transform(WT) and least squares support vector machine(LS-SVM) was applied to forecast annual the maximum earthquake magnitude in the South China seismic area. Firstly, using wavelet transform the series of annual the maximum earthquake magnitude in the South China seismic area was decomposed into several sub-sequences on different scales(frequency band). Secondly, the least squares support vector machine was applied to forecast the decomposed sub-sequences, respectively. Finally, the reconstruction of forecasted sub-sequences was used as the final predicted result. The comparison with LS-SVM and period method was made. The forecasted results show that WT and LS-SVM method possesses higher prediction precision and excellent forecasting effect, and prove that the method to combine WT with LS-SVM is feasible to forecast time series of strong earthquakes in the South China seismic area. The forecasted results also suggest that there are the strong and different nonlinear relations between variations of velocity of the Earth's rotation with high frequency portion (trend) and low frequency portion(short variations) of time sequence of the annual maximum earthquake in the South China seismic area.
引文
崔万照、朱长纯、保文星等,2004,混沌时间序列的支持向量机预测,物理学报,53(10),3304～3309。
飞思产品研发中心,2005,小波分析理论与MATLAB7实现,北京:电子工业出版社。
裴韬、周成虎、汪闽等,用二进小波分析方法对华北地区强震活动期的研究,地震研究,27(11),37～42。
傅征祥,1981,浅源强震分布与地球自转速率变化的关系,地震科学研究,(2),7～12。
傅征祥、邵辉成、丁香,2004,中国大陆浅源强震分布与地球自转变化的关系,地震,24(3),15～20。
贾晋康、王仁,1981,地球自转速率变化与地震触发,地震科学研究,(3),1～8。
宋超、黄民翔、叶剑斌,2002,小波分析方法在电力系统短期负荷预测中的应用,电力系统及其自动化学报,11(3),8～12。
孙德山、吴今培、侯振挺等,2004,基于SVR的混沌时间序列预测,计算机工程应用,54(2),54～56。
王景雷、吴景社、孙景生等,2003,支持向量机在地下水位预测中的应用研究,水利学报,49(5),124～128。
王炜、刘悦、李国正等,2005,中国大陆强震时间序列预测的支持向量机方法,地震,25(4),26～32。
许建华、张学工、李衍达,2004,支持向量机的新发展,控制与决策,19(5),481～485。
徐军华、刘天琪,2004,基于小波分解和人工神经网络的短期负荷预测,电网技术,28(8),30～33。
叶美盈、汪晓东、张浩然,2005,基于在线最小二乘支持向量机回归的混沌时间序列预测,物理学报,54(6),2568～2573。
宇缨、李清华,2005,统计学习理论和支持向量机,沈阳大学学报,17(4),42～47。
Cristanini,N.,ShaweTaylor,J.,2004,支持向量机导论,李国正、王猛、曾华军译,北京,电子工业出版社。
VapnikV.,2004,统计学习理论,许建华、张学工译,北京:电子工业出版社。
飞思产品研发中心,2005,小波分析理论与MATLAB7实现,北京:电子工业出版社。
裴韬、周成虎、汪闽等,用二进小波分析方法对华北地区强震活动期的研究,地震研究,27(11),37～42。
傅征祥,1981,浅源强震分布与地球自转速率变化的关系,地震科学研究,(2),7～12。
傅征祥、邵辉成、丁香,2004,中国大陆浅源强震分布与地球自转变化的关系,地震,24(3),15～20。
贾晋康、王仁,1981,地球自转速率变化与地震触发,地震科学研究,(3),1～8。
宋超、黄民翔、叶剑斌,2002,小波分析方法在电力系统短期负荷预测中的应用,电力系统及其自动化学报,11(3),8～12。
孙德山、吴今培、侯振挺等,2004,基于SVR的混沌时间序列预测,计算机工程应用,54(2),54～56。
王景雷、吴景社、孙景生等,2003,支持向量机在地下水位预测中的应用研究,水利学报,49(5),124～128。
王炜、刘悦、李国正等,2005,中国大陆强震时间序列预测的支持向量机方法,地震,25(4),26～32。
许建华、张学工、李衍达,2004,支持向量机的新发展,控制与决策,19(5),481～485。
徐军华、刘天琪,2004,基于小波分解和人工神经网络的短期负荷预测,电网技术,28(8),30～33。
叶美盈、汪晓东、张浩然,2005,基于在线最小二乘支持向量机回归的混沌时间序列预测,物理学报,54(6),2568～2573。
宇缨、李清华,2005,统计学习理论和支持向量机,沈阳大学学报,17(4),42～47。
Cristanini,N.,ShaweTaylor,J.,2004,支持向量机导论,李国正、王猛、曾华军译,北京,电子工业出版社。
VapnikV.,2004,统计学习理论,许建华、张学工译,北京:电子工业出版社。
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