数学物理学报(英文版) ›› 1995, Vol. 15 ›› Issue (S1): 123-132.

• Articles • 上一篇    下一篇

STRONG EMBEDDING OF PRODUCT-LIMIT ESTIMATOR OF BIVARIATE SURVIVAL DISTRIBUTION FUNCTION UNDER RANDOM CENSORSHIP

王启华   

  1. Department of Probability and statistics, Peking University, Beijing 100871, China
  • 收稿日期:1993-03-09 出版日期:1995-12-31 发布日期:1995-12-31
  • 基金资助:
    Supported by the doctoral programme scientific foundation and NSFC.

STRONG EMBEDDING OF PRODUCT-LIMIT ESTIMATOR OF BIVARIATE SURVIVAL DISTRIBUTION FUNCTION UNDER RANDOM CENSORSHIP

Wang Qihua   

  1. Department of Probability and statistics, Peking University, Beijing 100871, China
  • Received:1993-03-09 Online:1995-12-31 Published:1995-12-31
  • Supported by:
    Supported by the doctoral programme scientific foundation and NSFC.

摘要: Let (Xi0,Yi0), 1 ≤ i ≤ n be i. i. d nonnegative random vectors with continuous survival distribution function be the product-limit estimator of S(s,t)=P(X0 > s,Y0 > t). Let Sn(s,t) be the product-limit estimator of S(s,t) suggested by Campbell and Földes (1980). In this paper it is shown that under some conditions a sequence of Gaussian processes Gn(s,t) can be constructed such that
sup0T0S|n(1)/2[Sn(s,t)-S(s,t)]-Gn(s,t)|=O(n-(1)/2 log2n) a. s.,for S,T which together satisfy a certain condition.

关键词: Strong Embedding, Product-Limit Estimator, Survival Distribution

Abstract: Let (Xi0,Yi0), 1 ≤ i ≤ n be i. i. d nonnegative random vectors with continuous survival distribution function be the product-limit estimator of S(s,t)=P(X0 > s,Y0 > t). Let Sn(s,t) be the product-limit estimator of S(s,t) suggested by Campbell and Földes (1980). In this paper it is shown that under some conditions a sequence of Gaussian processes Gn(s,t) can be constructed such that
sup0T0S|n(1)/2[Sn(s,t)-S(s,t)]-Gn(s,t)|=O(n-(1)/2 log2n) a. s.,for S,T which together satisfy a certain condition.

Key words: Strong Embedding, Product-Limit Estimator, Survival Distribution