| Popis: |
For testing normality with unknown parameters μ and σ2, extensive simulation studies have shown the superiority of the Shapiro–Wilk test and variants. The Shapiro–Francia test is a simplified version of the Shapiro–Wilk test, which uses the squared normal probability plot correlation as a test statistic. The Cramer–von Mises and Anderson–Darling tests with plug-in estimators for μ and σ2 perform slightly worse than the Shapiro–Wilk test. Keywords: normality tests; Shapiro–Wilk test; Shapiro–Francia test; Ryan–Joiner test; Lilliefors test; Anderson–Darling test; Cramer–von Mises test; Jarque–Bera test |
