The Use of $\omega ^2 $ Tests for Testing Parametric Hypotheses
The Use of $\omega ^2 $ Tests for Testing Parametric Hypotheses
复制标题
使用 $omega ^2 $ 检验来检验参数假设
DOI:
10.1137/1124035
复制
发表时间:
1980
期刊:
影响因子:
--
通讯作者:
E. Khmaladze
中科院分区:
文献类型:
--
作者:
E. Khmaladze
As we know, the (. 0 2 test for testing a simple hypothesis about an unknown continuous distribution function (df) F (x) of independent identically distributed random variables (rv’s) X1,’’’, X,, is based on the statistic to, n [F,(x)-Fo (x)] 2 dFo (x), where F,(x) denotes the empirical distribution function (edf) of the rv’s X1, X2,’, X,; according to the hypothesis, F (x) Fo (x). Information about the to 2 test can be found, for example, in [1] or [2]. Carrying out the usual substitution Fo (x), we obtain o,= n [F,(t)-t] dt= v,(t) dr, where F,(t) denotes the edf, of the independent uniformly distributed rv’s T= Fo (X),..., T= Fo (X), and v (t)=[F (t)-t]. The sequence of dis-tributions P o the processes v (t) converges weakly in L [0, 1] as n m to the distribution P of the Brownian bridge v (t), and consequently the distribution of the statistic converges weakly to the distribution ofthe rv’s. v (t) dt.