A conservative test and confidence region for comparing heteroscedastic regressions (Q3774752)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 4034900
Language Label Description Also known as
default for all languages
No label defined
    English
    A conservative test and confidence region for comparing heteroscedastic regressions
    scientific article; zbMATH DE number 4034900

      Statements

      A conservative test and confidence region for comparing heteroscedastic regressions (English)
      0 references
      0 references
      0 references
      0 references
      1988
      0 references
      comparing heteroscedastic regressions
      0 references
      strong unimodality
      0 references
      extension of the Behrens-Fisher problem
      0 references
      comparison of two regression planes
      0 references
      unequal residual variances
      0 references
      Studentized maximum and minimum chi-squared distribution
      0 references
      conservative tests
      0 references
      confidence regions
      0 references
      log concave distribution function of the error terms
      0 references
      In the paper a practically important extension of the Behrens-Fisher problem of comparing the means of two normal populations with unequal variances is considered. The extension is the comparison of two regression planes with possibly unequal residual variances in the two corresponding linear regression models. Both models are assumed to have the same known design matrix. Then an analogoue of the Behrens-Fisher statistic is proposed and its distribution bounded in terms of the Studentized maximum and minimum chi-squared distribution. These results are used to construct conservative tests and confidence regions for comparing the regression parameters of both models for an arbitrary log concave distribution function of the error terms. Computational comparisons between the new methods and earlier methods indicate few discrepancies but the earlier methods are applicable only under restrictive conditions.
      0 references

      Identifiers