摘要
地震是一种非线性现象 ,因而很多人计算地震分布的分形维数 ,但具体各种计算方法的误差 (或精度 )是多少 ,还没有定量的估计。鉴于地震空间分布具有有限、离散、点集的特点 ,用双标度 Contor多分形集理论模型数值模拟来估算其精度 (误差 ) ,并判定各种方法的优劣。理论模型数值模拟得出如下结论 :1随着样本容量的增大 ,计算精度会提高 ;2固定半径法 ( RAD)计算误差偏大 ,固定质量法 ( MAS)和最小生成树法 ( MST)较好 ;3当样本容量达到约 2 0 0时 ,MAS法和 MST法计算误差大体可稳定在 0 .0 5的范围内。
Earthquake is a non line phenomenon, so many seismologists calculate the fractal dimension of seismic distribution. However, the precision of calculation method is uncertain. Because the space time distribution of earthquakes is characteristics of limitation, dispersion and point sets, we use the digital simulation of theory models of two dimension double rule Contor sets to estimate its precision (or error), and to judge some methods good or not. From the theory models simulated digitally, we obtained some results. ① As the samples increasing, the precision of several methods is increasing. ② The error of the fixed radius method (RAD) is too big, while ones of the fixed mass method (MAS) and the minimal spanning tree method (MST) is smaller. ③ When the numbers of samples are more than 200, the calculation precision of MAS and MST is around 0.05.
出处
《地震》
CSCD
北大核心
2000年第3期1-8,共8页
Earthquake
基金
地震科学联合基金会资助项目!( 92 2 87)
关键词
多重分形
精度
地震
离散
点集
空间分布
Multi fractal
Precision
Theory model
Numbers of samples
