pone.0233901.g003.tif (1.26 MB)
Download filePower of the introduced test and standard tests for normality for different sample sizes: N = 50, N = 100; N = 200 and N = 1000 for the generalized Gaussian distribution with respect to the excess kurtosis.
figure
posted on 11.06.2020, 20:06 by Agnieszka Wyłomańska, D. Robert Iskander, Krzysztof BurneckiPowers were calculated on the basis of 5000 simulations. The significance level is equal to 5%.
History
Usage metrics
Read the peer-reviewed publication
Categories
Keywords
hypothesis teststest statisticprobability distributionnovel approachsignal processing techniquesdata examplesleptokurtic onesplatykurtic distributionstesting normalityEdgeworth expansion Statistical inferenceOmnibus testEdgeworth expansionsample sizesconfidence intervalstime seriesleptokurtic distributions
CC BY 4.0