The likelihood ratio test for detecting the best treatment among several exponential populations

지수분표에 있어서 최우수 처리의 판별을 위한 우도비 검정

  • 황형태 ((140-714) 서울시 용산구 한남동 산8, 단국대학교 전산통계학과)
  • Published : 1995.03.01

Abstract

The method for detecting the best treatment is considered by means of hypothesis testing in the exponential case. The likelihood ratio test for a given hypothesis is derived to control the error probability, and the minimum powers in the interested regions are calculated to design the sampling plan.

다수의 지수분포에 있어서 최우수 처리를 판별하는 방법을 가설검정의 형식으로 연구하였다. 최우수 처리의 판별을 위해 적절한 가설을 설정하고, 주어진 유의수준을 만족시키는 우도비 검정 방법을 유도하였으며, 관심영역에서의 최소 검정력을 계산함으로써 실제 표본 설계가 가능하도록 하였다.

Keywords

References

  1. Annals of Mathematical Statistics v.25 A single sample multiple decision procedure for ranking means of normal populations with known variances Bechhofer,R.E.
  2. Ph.D. Thesis(Mimeo. Ser. No. 150), Institute of Statistics, Univ. of North Carolina On a decision rule for a problem in ranking means Gupta,S.S.
  3. Annals of Statistics v.12 Constrained simultaneous confidence intervals for multiple comparisons with the best Hsu,J.C.
  4. Design of Experiments:Ranking and Selection Ranking, selection and multiple comparisons with the best Hsu,J.C.
  5. Journal of the Korean Statistical Society v.17 On detecting the best treatment Kim,W.C.
  6. Journal of the Korean Statistical Society v.20 Empirical Bayesian multiple comparisons with the best Kim,W.C.;Hwang,H.T.