Journal of applied mathematics & informatics

The Korean Society for Computational and Applied Mathematics (한국전산응용수학회)
권호 표지

INFLUENCE ANALYSIS FOR A LINEAR HYPOTHESIS IN MULTIVARIATE REGRESSION MODEL

  • Published : 2003.09.01

⟨ Previous Next ⟩

Abstract

The influence of observations on the Wilks' lambda test of a linear hypothesis in multivariate regression is investigated using the local influence method. The perturbation scheme of case-weights is considered. A numerical example is given to show the effectiveness of the local influence method in identifying the influential observations.

Keywords

References

  1. J. Amer. Statist v.87 General classes of influence measures for multivariate regression B.E.Barrett;R.F.Ling
  2. J. Roy. Statist. Soc. Sec. B v.48 Assessment of local influence(with discussion) R.D.Cook
  3. em J. Appl. Statist. v.16 Detection of influential observations in multivariate regression A.Hossain;D.N.Naik
  4. Commun. Statist.-Theory Meth. v.24 Local influence in multivariate regression M.G.Kim
  5. J. Appl. Statist. v.27 Outliers and influential observations in the structural errors-in-variables model M.G.Kim
  6. J. Appl. Math. & Computing v.10 Case-deletion diagnostics for testing a linear hypothesis about regression coefficients M.G.Kim
  7. Multivariate Observations G.A.F.Seber
  8. Sankhya B v.58 Local inflence of test statistics in multivariate regression M.K.Tang;W.K.Fung
  9. J. Roy. Statist. Soc. Sec. B v.55 Second-order approach to local influence X.Wu;Z.Luo