在左截断数据下构造了回归函数的变窗宽局部线性M估计,得到了该估计的弱相合性以及渐近正态性结果,把文[1]在完全数据下的结果推广到左截断数据下.模拟结果表明该估计处理离群数据时具有稳健性.
Abstract: This paper constructed a variable bandwidth and local linear M-estimator of regression function under left truncated data, obtained its weak consistency and asymptotic normality, and extended the results for complete data in document [1] to left-truncated data. The results show the estimator has robustness when it deals with outlier data.