摘要
实际测量中存在数据误差分布的偏差和异常点问题,然而经典的2k+1点移动光滑方法不具有稳定性,为解决这一问题,采用M稳健二次多项式线性回归空间回归方法进行稳健光滑,通过对影响函数的分析,指出该方法具有全稳健性能的优点。
There exist deviations of error distribution and exceptional dots in practical data measurement. However the classical methods of shift smooth processing of 2k ＋ 1 dots are not robust. In order to deal with this problem, the paper utilized M robust quadratic polynomial linearity space regression algorithm to deal with the robust smooth progressing. Lastly, the full robustness of the algorithm is demonstrated by analyzing the IF,
出处
《计算机应用》
CSCD
北大核心
2008年第7期1678-1680,1695,共4页
journal of Computer Applications
关键词
数据光滑
M稳健回归
影响函数
data smooth processing
M robust regression
influence function
