Detection of the Normal Population with the Largest Absolute Value of Mean

  • Kim, Woo-Chul (Department of Computer Science and Statistics, Seoul National University, Seoul, 151-742) ;
  • Jeong, Gyu-Jin (Department of Applied Statistics, Hannam University, Daejeon 300-791)
  • Published : 1993.06.01

Abstract

Among k independent normal populations with unknown means and a common unknown variance, the problem of detecting the population with the largest absolute value of mean is considered. This problem is formulated in a manner close to the framework of testing hypotheses, and the maximum error probability and the minimum power are considered. The power charts necessary to determine the sample size are provided. The problem of detecting the population with the smallest absolute value of mean is also considered.

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. Communication in Statistics, -Theory and Methods v.A17 The least significant spacing for one versus the rest normal populations Bofinger,E.
  3. Ph.D. Thesis (Mimeo. Ser No. 150), Institute of Statistics, University of North Carolina On a decision rule for a problem in ranking means Gupta,S.S.
  4. Communications in Statistics v.B14 On the distribution of the Studentized maximum of equally correlated normal random variables Gupta,S.S.;Panchapakesan,S.;Sohn,J.K.
  5. Unpublished manuscript, Northesatern University Is the selected population unsurpassed? (Three or four normal populations) Gutmann,S.
  6. Annals of Statistics v.15 Is the selected population the best? Gutmann,S.;Maymin,Z.
  7. Journal of The Korean Society for Quality Control v.16 no.2 Confidence bounds for superiority Jeon,J.W.;Kim,W.C.;Jeong,G.J.
  8. Journal of The Korean Statistical Society v.17 On detecting the best treatment Kim,W.C.
  9. Technometrics v.13 Some selection problems involving folded normal distribution Rizvi,M.H.