论文部分内容阅读
目前时间序列分析在工业及水文、气象、地球物理等自然科学领域中得到了迅速发展。本文用非线性门限自回归及最大熵谱分析方法分析了我国年最大震级和全球年地震能量释放序列。最大熵谱分析具有分辨率高、周期偏移小的突出优点。对1897年至1983年的我国最大震级和全球年地震能量序列进行了最大熵谱分析,计算结果表明,我国年最大震级的周期为3.9年,全球年地震能量的显著周期为7.7年、4.4年、2.9年、及2.3年。对上述序列进行了门限自回归分析,并对我国分区年最大震级序列进行了门限自回归分析。在门限延迟值d=4,8,10时相应的AIC值为谷值。
At present, time series analysis has been rapidly developed in the fields of industry, hydrology, meteorology and geophysics. In this paper, we use non-linear threshold autoregression and maximum entropy spectrum analysis method to analyze the annual maximum magnitude and global annual earthquakes energy release sequence. Maximum entropy spectral analysis with high resolution, small periodic offset advantages. The maximum entropy spectrum analysis of the maximum earthquake magnitude and the annual global earthquake energy series from 1897 to 1983 shows that the annual maximum earthquake magnitude is 3.9 years and the annual global earthquake energy significant period is 7.7 years and 4.4 years , 2.9 years, and 2.3 years. Threshold autoregressive analysis was carried out on the above sequence, and the threshold autoregressive analysis of the maximum magnitude sequence in China was conducted. When the threshold delay value d = 4, 8, and 10, the corresponding AIC value is the valley value.