右删失左截断情形下分布函数的分位数估计

THE PRODUCT-LIMIT QUANTILE ESTIMATOR FOR RANDOMLY TRUNCATED AND CENSORED DATA

文中考虑了右删失左截断数据情形下分布函数的分位数估计，讨论了该估计的渐近性质并获得了它的强弱Bahadur类型的表示定理．利用此Bahadur表示定理很容易获得该分位数估计的渐近正态性及置信区间等结果．

篎or the product-limit(PL)-estimator Fn in [6] the quantile PL-estimator np= (p) = in f {t: Fn(t) ≥p} is proposed as an estimator of quantile function F-1(P), 0< p<1 baized on left truncated and right censored data. Some Bahadur type representations for this PL quantile estimator are established. The advantage of these results are illustrated via obtaining confidence intervals and band for quantiles of F.