地质变量的稳健修匀

ROBUST FILTERING OF GEOLOGICAL VARIABLES

本文从理论上并用蒙特卡洛方法分析比较了游动中值、游动平均的修匀效果及优缺点,据此研究了基于Huber-M估计的稳健修匀方法,理论分析,实例计算、蒙特卡洛模拟结果表明:它既有较强的稳健性,又有较好的效率,还有较高的抑制随机误差能力,当地质变量存在异常时,这是一种可以替代游动中值,游动平均的较理想的修匀方法。

The filtering effect of the moving mean and moving median value is discussed theoratically and simulated by the method of Monte Carlo. Then a robust filtering method based on Huber M-estimate is studied. The results of theoratical analysis, practical calculation, Monte Carlo simulation indicate that the method is highly robust and efficient, and can restrain the random error. It is a useful filtering method instead of the moving mean and moving median value if outliers exist in geological variables.