Journal of the Korean Statistical Society

The Korean Statistical Society (한국통계학회)
권호 표지

Asymptotic Relative Efficiency of Chi-squared Type Tests Based on the Empirical Process

  • Lee, Sang-Yeol (Department of Statistics, Sookmyung Women's University, Seoul, 140-742, Korea.)
  • Published : 1996.09.01
Download PDF
⟨ Previous Next ⟩

Abstract

The chi-squared type statistic generated from the empirical process can be used for testing the goodness of fit hypothesis on iid random sample. Lee (1995) showed that under some conditions, the chi-squared type statistic is asymptotically maximin in the sense of Strasser (1985). Since the chi-squared type statistic depends on the choice of *points in the unit interval, it is worth investigating the points yielding more efficient tests. Motivated by this viewpoint, we are led to study the asymptotic relative efficiency of chi-squared type tests in the same setting of Lee (1995). Some examples are given for illustration.

Keywords

References

  1. Convergence of Probability Measures Billingsley, P.
  2. Annals of Statististics v.14 Testing for normality in arbitrary dimension Csorgo, S.
  3. Annals of Statistics v.1 Weak vonvergence of the sample distribution function when parameters and estimated Durbin, J.
  4. Journal of Royal Statistical Society, B v.43 A goodness of fit test based on the empirical characteristic function when parameters must by estimated Koutrouvelis, I. A.;Kellermeire, J.
  5. To appear in Statistics and Probability Letters Asymptotic maximin property of chi-squared type tests based on the empirical process Lee, S.
  6. Empirical Processes with Applications to Statistics Shorack, G.;Wellner, J.
  7. Mathematical Theory of Statistics Strasser, H.