Academia.eduAcademia.edu

Outline

Title
Introduction
Invariance of the Distributions of Test Statistics
References
All Topics
Mathematics
Statistics

Some new tests for multivariate normality

Profile image of Adolfo José QuirozAdolfo José Quiroz

1991, Probability Theory and Related Fields

Related papers

Journal of Statistical Computation and Simulation A MATLAB package for multivariate normality test A MATLAB package for multivariate normality test
Mohammad Irandoust

A MATLAB package testing for multivariate normality (TMVN) is implemented as an interactive and graphical tool to examine multivariate normality (MVN). Monte Carlo simulation studies have failed to find a uniformly most powerful MVN test, which requires a rather extensive statistical inference procedure. TMVN contains several competitive MVN tests and provides a flexible and extensive testing environment for univariate or multivariate data analyses. Simulated results provide information of which test may possess more power for the selected non-MVN alternatives. Fisher's Iris data are used to show how TMVN can be used in practice.

View PDFchevron_right
An Extensive Comparisons of 50 Univariate Goodness-of-fit Tests for Normality
Stanislaus S . Uyanto

Austrian Journal of Statistics

The assumption of normality needs to be checked for many statistical procedures, namely parametric tests, because their validity depends on it. Given the importance of this subject and the widespread development of normality tests, comprehensive descriptionsand power comparisons of such tests are of considerable interest. Since recent comparison studies do not include several interesting and more recently developed tests, a further comparison of normality tests is considered to be of foremost interest. This study addresses the performance of 50 normality tests available in literature, from 1900 until 2018. Because a theoretical comparison is not possible, Monte Carlo simulation were used from various symmetric and asymmetric distributions for different sample sizes ranging from 10 to 100. The simulations results show that for symmetric distributions with support on (−∞, ∞) the tests Robust Jarque–Bera and Gel–Miao–Gastwirth have generally the most power. For asymmetric distributions...

View PDFchevron_right
Evaluating Multivariate Normality: A Graphical Approach
Shumila Ramzan

2013

The statistical graphics play an important role in providing the insights about data in the process of data analysis. The main objective of this paper is to provide a comprehensive review of the methods for checking the normality assumption. Multivariate normality is one of the basic assumptions in multivariate data analysis. Univariate normality is essential for the data to be multivariate normal. This paper reviews graphical methods for evaluating univariate and multivariate normality. These methods are applied on a real life data set and the normality is investigated.

View PDFchevron_right
Ways To Evaluate the Assumption of Multivariate Normality
Alyce Ashcraft

1998

This paper reviews graphical and nongraphical methods for estimating multivariate normality. Prior to exploring this methodology, a foundation is established by presenting ways to assess univariate and bivariate normality. A data set of three variables used by J. Stevens (1986) is analyzed using Q-Q plots, stem and leaf plots, histograms, skewness, and kurtosis coefficients, the Shapiro-Wilk statistic, and bivariate and multivariate scatterplots. Multivariate normality is explored in terms of calculating Mahalanobis distances and plotting them on a scattergram against derived chi-square values using Fortran and Statistical Package for the Social Sciences (SPSS) programs developed by B. Thompson (1990Thompson ( , 1997. Appendixes, which comprise more than half the half, contain the SPSS commands, two computer programs for the analysis, and some results of the analyses. (Contains 24 figures and 11 references.)

View PDFchevron_right
Evaluating Univariate, Bivariate, and Multivariate Normality Using Graphical Procedures
Thomas K Burdenski

2002

This paper reviews graphical and nongraphical procedures for evaluating multivariate normality by guiding the reader through univariate and oivariate procedures that are necessary, but insufficient, indications 0± a multivariate normal distribution. A data set using three dependent variables for two groups provided by D. George and P. Mallery (1999) is used to analyze histograms, stem-and-leaf plots, box-and-whisker plots, kurtosis and skewness coefficients, Q-Q plots, the Shapiro-Wilk or Kolmogorov-Smirnov statistic, and bivariate scatterplots. A procedure programmed by B. Thompson

View PDFchevron_right
A normality test for multivariate dependent samples
sara elbouch

Signal Processing

Most normality tests in the literature are performed for scalar and independent samples. Thus, they become unreliable when applied to colored processes, hampering their use in realistic scenarios. We focus on Mardia's multivariate kurtosis, derive closed-form expressions of its asymptotic distribution for statistically dependent samples, under the null hypothesis of normality. Included experiments illustrate, by means of copulas, that it does not suffice to test a one-dimensional marginal to conclude normality. The proposed test also exhibits good properties on other typical scenarios, such as the detection of a non-Gaussian process in the presence of an additive Gaussian noise.

View PDFchevron_right
Mardia’s Skewness and Kurtosis for Assessing Normality Assumption in Multivariate Regression
Dewi wulandari

Enthusiastic : International Journal of Applied Statistics and Data Science, 2021

In Multivariate regression, we need to assess normality assumption simultaneously, not univariately. Univariate normal distribution does not guarantee the occurrence of multivariate normal distribution [1]. So we need to extend the assessment of univariate normal distribution into multivariate methods. One extended method is skewness and kurtosis as proposed by Mardia [2]. In this paper, we introduce the method, present the procedure of this method, and show how to examine normality assumption in multivariate regression study case using this method and expose the use of statistics software to help us in numerical calculation. Received February 20, 2021Revised March 8, 2021Accepted March 10, 2021

View PDFchevron_right
A Brief Review of Tests for Normality
Keya R A N I Das

American Journal of Theoretical and Applied Statistics, 2016

In statistics it is conventional to assume that the observations are normal. The entire statistical framework is grounded on this assumption and if this assumption is violated the inference breaks down. For this reason it is essential to check or test this assumption before any statistical analysis of data. In this paper we provide a brief review of commonly used tests for normality. We present both graphical and analytical tests here. Normality tests in regression and experimental design suffer from supernormality. We also address this issue in this paper and present some tests which can successfully handle this problem.

View PDFchevron_right
Empirical Investigation of Type 1 Error Rate of Univariate Tests of Normality
Kayode Ayinde

International Journal of Computer Applications, 2016

Normality assumption is important in univariate parametric statistical tests. Either the varibles or the error terms in the model have to be normally distributed before valid statistical conclusions could be made. Various tests of univariate normality including that of Pearson,

View PDFchevron_right
A test for multivariate skew-normality based on its canonical form
Bruno Scarpa

Journal of Multivariate Analysis, 2014

A test to assess if a sample comes from a multivariate skew-normal distribution is proposed. The test statistic is obtained from the canonical form of the multivariate skew-normal distribution and its null distribution is derived. The power of the proposed test is evaluated through Monte Carlo simulations for different conveniently chosen alternatives. Finally, three numerical examples are presented for the purpose of illustration.

View PDFchevron_right

References (37)

  1. Alexander, K.S.: The central limit theorem for empirical processes on Vapnik-Cervonenkis classes. Ann. Probab. !5, 178 203 (1987)
  2. Anderson, T.W., Stephens, M.A.: Tests for randomness of directions against equatorial and bimodal alternatives. Biometrika 59, 613 621 (1972)
  3. Andrews, D.W.K.: Chi-square diagnostic tests for econometric models: introduction and appli- cations. J. Econ. 37, 135 156 (1988)
  4. Barndorff-Nielsen, O.: Information and exponential families in statistical theory. New York: Wiley 1978
  5. Bera, A., John, S.: Tests for multivariate normality with Pearson alternatives. Commun. Stat. Theory Methods A12, 103-117 (1983)
  6. Beran, R.J.: Testing for uniformity on a compact homogeneous space. J. Appl. Probab. 5, 177-195 (1968)
  7. Brown, L.D.: Fundamentals of statistical exponential families. Inst. of Math. Statist. Lecture Notes -Monograph Series 9 (1986)
  8. Chernoff, H., Lehmann, E.L.: The use of maximum likelihood estimates in Z 2 tests for goodness of fit. Ann. Math. Statist. 25, 579-586 (1954)
  9. Cstrgt, Sandor: Testing for normality in arbitrary dimension. Ann. Statist. 14, 708-723 (1986) Dieudonnt, J.: Foundations of modern analysis. New York: Academic Press 1960
  10. Dudley, R.M. : On )~z tests of composite hypotheses. Probability Theory, Banach Center Publica- tions 5, 75 87. Warsaw: PWN 1979
  11. Dudley, R.M. : A course on empirical processes. In: Hennequin, P.L. (ed.) Ecole d'tt6 de probabi- litts de St.-Flour, 1982. (Lect. Notes Math., vol. 1097, pp. vii viii, 1-142) Berlin Heidelberg New York: Springer 1984
  12. Dudley, R.M.: An extended Wichura theorem, definitions of Donsker class, and weighted empirical distributions. Probability in Banach Spaces V (Proc. Conf. Medford, 1984; Lect. Notes Math., vol. 1153, pp. 141-178) Berlin Heidelberg New York Tokyo: Springer 1985
  13. Dudley, R.M.: Universal Donsker classes and metric entropy. Ann. Probab. 15, 1306-1326 (1987)
  14. Dudley, R.M., Philipp, W.: Invariance principles for sums of Banach space valued random elements and empirical processes. Z. Wahrscheinlichkeitstheor. Verw. Geb. 62, 509 552 (1983)
  15. Dzhaparidze, K.O., Nikulin, M.S.: On a modification of the standard statistics of Pearson. Theory Probab. Appl. 19, 851 853 (1974)
  16. Fernholz, L.T.: von Mises calculus for statistical functionals. (Lect. Notes Statist., vol. 19) Berlin Heidelberg New York: Springer 1983
  17. Gint, E.: Invariant tests for uniformity on compact Riemannian manifolds based on Sobolev norms. Ann. Statist. 3, 1243-1266 (1975)
  18. Holt, R.: Computation of gamma and beta tail probabilities. Mathematics Dept., Massachusetts Institut of Technology 1986
  19. Huber, P.J.: The behavior of maximum likelihood estimates under nonstandard conditions. In: Proc. 5th Berkeley Syrup. Math. Statist. Prob. 1965, vol. 1, pp. 221-233. Berkeley Los Angeles: University of California Press 1967
  20. Joiner, B.L., Rosenblatt, J.R.: Some properties of the range in samples from Tukey's symmetric lambda distributions. J. Am. Statist. Assoc. 66, 394-399 (1971)
  21. Koziol, J.A.: On assessing multivariate normality. J. R. Stat. Soc., Ser. B45, 358-361 (1983)
  22. Lehmann, E.L.: Testing statistical hypotheses, 2d edn. New York: Wiley 1986
  23. Mardia, K.V.: Measures of multivariate skewness and kurtosis with applications. Biometrika 57, 519-530 (1970)
  24. Moore, D.S.: A chi-square statistic with random cell boundaries. Ann. Math. Statist. 42, 147-156 (1971)
  25. Moore, D.S., Stubblebine, J.B.: Chi-square tests for multivariate normality with application to common stock prices. Commun. Stat. Theory Methods A10, 713-738 (1981)
  26. Mfiller, C.: Spherical harmonics (Lect. Notes Math., vol. 17). Berlin Heidelberg New York: Springer 1966
  27. Pollard, D.: General chi-square goodness-of-fit tests with data-dependent cells. Z. Wahrschein- lichkeitstheor. Verw. Geb. 50, 317-331 (1979)
  28. Pollard, D.: A central limit theorem for empirical processes. J. Aust. Math. Soc. Set. A33, 235248 (1982)
  29. Pollard, D.: Convergence of stochastic processes. New York Berlin Heidelberg: Springer 1984
  30. Prentice, M.J.: On invariant tests of uniformity for directions and orientations. Ann. Statist. 6, 169 176 (1978)
  31. Quiroz, A.J.: On Donsker classes of functions and their application to tests for goodness of fit. Ph. D. thesis, mathematics, Mass. Inst. of Tech. (1986)
  32. Rao, K.C., Robson, D.S.: A chi-square statistic for goodness-of-fit tests within the exponential family. Comm. Statist. 3, 1139-1153 (1974)
  33. Roy, A.R. : On Z 2 statistics with variable intervals. Technical Report no. i, Dept. of Statistics, Stanford University (1956)
  34. Small, N.J.H. : Plotting squared radii. Biometrika 68, 657-658 (1978)
  35. Stein, E.M., Weiss, G. : Introduction to Fourier analysis on Euclidean Spaces. Princeton: Prince- ton University Press 1971
  36. Watson, G.S.: On chi-square goodness of-fit tests for continuous distributions. J. R. Star. Soc., Ser. B20, 44-61, with discussion 61 72 (1958)
  37. Wenocur, R.S., Dudley, R.M.: Some special Vapnik-Chervonenkis classes. Discrete Math. 33, 313-318 (1981)

Related papers

A new goodness of fit test for multivariate normality
Orhan Kesemen

Hacettepe Journal of Mathematics and Statistics, 2021

This paper presents a multivariate Kolmogorov-Smirnov (MVKS) goodness of fit test for multivariate normality. The proposed test is based on the difference between the empirical distribution function and the theoretical distribution function. While calculating them in multivariate case, the problem is that the variables cannot be distribution-free as in the univariate case. Firstly, the variables are made independent to solve this problem and the Rosenblatt transform is applied for independence of variates. Then the theoretical and empirical distribution values are calculated and the MVKS test statistic is computed. It provides an easy calculation for d-dimensional data by using the same algorithm and critical table values. This paper demonstrates the effectiveness of the MVKS for different dimensions with a simulation study which also includes the comparison of the MVKS critical tables with univariate Kolmogorov-Smirnov (KS) critical table and the power comparisons of the MVKS (bivariate case) against with the existing bivariate normality tests. Lastly, the MVKS is applied to two different multivariate data sets to confirm that it achieves consistent, accurate and correct results.

View PDFchevron_right
A New Jarque-Bera Type Omnibus Goodness-of-fit Test for Multivariate Normality
Mbanefo Madukaife

Journal of Advances in Mathematics and Computer Science

In this paper, a new Jarque-Bera type statistic for assessing the multivariate normality of a multivariate datasets is obtained. The affine-invariant and consistent statistic is shown to follow, asymptotically, a chi-square distribution with 2 degrees of freedom. The critical values of the test were evaluated empirically through extensive simulation studies for different sample sizes and different random vector dimensions. Also, the empirical type-I-error rates and empirical powers of the proposed test were compared with some other well-known competing statistics in the literature. The results obtained showed that the proposed omnibus statistic is a powerful tool for assessing multivariate normality (MVN) of multivariate datasets.

View PDFchevron_right
An Assessment of the Performances of Several Univariate Tests of Normality
James Adefisoye
View PDFchevron_right
Assessing Univariate and Multivariate Normality, A Guide For Non-Statisticians
Felix B Oppong

Most parametric methods rely on the assumption of normality. Results obtained from these methods are more powerful compared to their non-parametric counterparts. However for valid inference, the assumptions underlying the use of these methods should be satisfied. Many published statistical articles that make use of the assumption of normality fail to guarantee it. Hence, quite a number of published statistical results are presented with errors. As a way to reduce this, various approaches used in assessing the assumption of normality are presented and illustrated in this paper. In assessing both univariate and multivariate normality, several methods have been proposed. In the univariate setting, the Q-Q plot, histogram, box plot, stem-and-leaf plot or dot plot are some graphical methods that can be used. Also, the properties of the normal distribution provide an alternative approach to assess normality. The Kolmogorov-Smirnov a starting point for assessing normality in the multivariate setting. A scatter plot for each pair of variables together with a Gamma plot (Chi-squared Q-Q plot) is used in assessing bivariate normality. For more than two variables, a Gamma plot can still be used to check the assumption of multivariate normality. Among the many test proposed for testing multivariate normality, Royston's and Mardia's tests are used more often and are implemented in many statistical packages. When the normality assumption is not justifiable, techniques for non-normal data can be used. Likewise, transformation to near normality is another alternative.

View PDFchevron_right
Performances of Several Univariate Tests of Normality: An Empirical Study
B M Golam Kibria

Journal of Biometrics & Biostatistics, 2016

The problem of testing for normality is fundamental in both theoretical and empirical statistical research. This paper compares the performances of eighteen normality tests available in literature. Since a theoretical comparison is not possible, MonteCarlo simulation were done from various symmetric and asymmetric distributions for different sample sizes ranging from 10 to 1000. The performance of the test statistics are compared based on empirical Type I error rate and power of the test. The simulations results show that the Kurtosis Test is the most powerful for symmetric data and Shapiro Wilk test is the most powerful for asymmetric data.

View PDFchevron_right

Related topics

  • Mathematics
  • Statistics
  • Multivariate Statistics
  • Multivariate Normal Distribution
  • Asymptotic distribution
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved