论文信息 - Transformation diagnostics for linear models

Transformation diagnostics for linear models

SUMMARY An approximate transformation model is used to present an integrated development of tests for transformable deviations from a normal-theory linear model. Particular emphasis is placed on Tukey's test for transformable nonadditivity and its relation to tests for variahce heteroscedasticity and error skewness. Numerical illustrations of the tests are given.

David V. Hinkley | D. Hinkley

[1] S. Weisberg,et al. Residuals and Influence in Regression , 1982 .

[2] F. J. Anscombe,et al. The Examination and Analysis of Residuals , 1963 .

[3] D. Pregibon. Goodness of Link Tests for Generalized Linear Models , 1980 .

[4] R. D. Cook,et al. Transformations and Influential Cases in Regression , 1983 .

[5] P. McCullagh,et al. Generalized Linear Models , 1984 .

[6] G. Box. An analysis of transformations (with discussion) , 1964 .

[7] F. J. Anscombe,et al. Examination of Residuals , 1961 .

[8] F. Yates. A Monte-Carlo trial on the behaviour of the non-additivity test with nonnormal data , 1972 .

[9] Ronald D. Snee,et al. Nonadditivity in a Two-Way Classification: Is it Interaction or Nonhomogeneous Variance? , 1982 .

[10] G. W. Snedecor. STATISTICAL METHODS , 1967 .