A modified goodness-of-fit test based on likelihood ratio for the skew-normal distribution

Potas N., Sarisoy E. E., Kara M.

Applied Mathematical Sciences, no.77-80, pp.3869-3887, 2014 (Scopus) identifier

  • Publication Type: Article / Article
  • Publication Date: 2014
  • Doi Number: 10.12988/ams.2014.45354
  • Journal Name: Applied Mathematical Sciences
  • Journal Indexes: Scopus, zbMATH
  • Page Numbers: pp.3869-3887
  • Keywords: Likelihood ratio, Modified goodness-of-fit test, Skew-normal distribution
  • Ankara University Affiliated: Yes

Abstract

© 2014 Nihan Potas, Emre E. Sarisoy and Mahmut Kara.This research aims to find powerful and convenient goodness of fit test for the skew-normal distribution. Classic EDF goodness-of fit tests are the Kolmogorov-Smirnov KS, Anderson-Darling A2, Crame´r-von Mises W2, Liao-Shimokawa Ln tests and the statistical tests based on likelihood ratio, e.g., the Zhang-Kolmogorov-Smirnov ZK, Zhang-Anderson-Darling ZA, Zhang-Crame´r-von Mises ZC and modified Liao-Shimokawa test MLn which is proposed, the critical values of whichare calculated for all, tabulated for skewed normal distribution and compared with various shape parameters, alternative sampling distributions and sizes by using the Monte-Carlo simulation technique. In most situations the modified Liao-Shimokawa test is the most powerful goodness-of-fit test for the skew-normal distribution compared with other methods.