MONTE CARLO COMPARISON OF FOUR NORMALITY TESTS USING DIFFERENT ENTROPY ESTIMATES

Abstract

The paper studies four entropy tests of normality of real valued observations using four statistics based on different entropy estimates from n i.i.d. observations. For the test size α = 0.05 and sample sizes 5 ≤ n ≤ 50, it presents critical values of the test statistics. It also compares the power and robustness of these four tests using Monte Carlo computations for sample sizes n = 10 and n = 20. Preferences between the tests based on these computations can be extrapolated for small sample sizes.

Keywords:

• Test of normality
• Entropy test
• Entropy estimator
• Table of critical values
• Test power
• Sensitivity contour

ACKNOWLEDGMENTS

The research in this paper was supported in part by the grants GACR 102/99/1137, DGES PB96–0635 and GV99-159-1-01.