成功率的序贯检验与检验后成功率的置信限

The Sequential Tests for Success Ratio and the Confidence Limits of Success Ratio after Testing

  • 摘要: 设{Xii≥1}是一独立同分布随机变量列,q=P(Xi=0)=1-P(Xi=1);其中q未知.设{An:1≤nn0}和{Rn:1≤nn0}是任二个数集,它们满足A1A2≤…≤An0,0<R1R2≤…≤Rn0AiRii=1,…,no-1)和An0=RnoRn0
    对于假设H0qq0(0<q0<1)我们考虑序贯检验△=(rd),其中r=min{nn1DnAnDnRn}:d=IDrRr,这里D_n=\sum_i=1^n X_i(n \geq 1).IA是集A的示性函数,“d=1”意味着拒绝H0,“d=0”意味着接受H0,在本文中基于数据(rDr)我们找到了q的某种意义下的最优置性下(上)限。

     

    Abstract: Let Xii≥1 be a sequence of i.i. d. r.v.’s satisfying q=P(Xi=0)=1-P(Xi=1),q is unknown. Suppose that An:1≤nn0 and Rn:1≤nn0 are any two sets of numbers satisfying A1A2≤…≤An0,0<R1R2≤…≤Rn0,AiRii=1,…,no-1) and An0=RnoRn0.
    For the hypothesis H0qq0(0<q0<1), We consider the sequential test △=(rd) in which r=min{nn1DnAnorDnRn, d=IDrRr, where D_n=\sum_i=1^n X_i(n \geq 1).IA is the indicator of set A, “d =1” means rejecting H0 and “d = 0” means accepting H0. In the present paper, we find out the optimal lower (upper) confidence limits for q based on data (rDr) in some sence

     

/

返回文章
返回