TY - JOUR
AU - de Rozas, Gloria Perez Sainz
PY - 2003/01/23
Y2 - 2024/05/19
TI - Using Mathematica to build Non-parametric Statistical Tables
JF - Journal of Statistical Software
JA - J. Stat. Soft.
VL - 8
IS - 4
SE - Articles
DO - 10.18637/jss.v008.i04
UR - https://www.jstatsoft.org/index.php/jss/article/view/v008i04
SP - 1 - 26
AB - In this paper, I present computational procedures to obtian statistical tables. The tables of the asymptotic distribution and the exact distribution of Kolmogorov-Smirnov statistic D<sub>n</sub> for one population, the table of the distribution of the runs R, the table of the distribution of Wilcoxon signed-rank statistic W<sup>+</sup> and the table of the distribution of Mann-Whitney statistic U<sub>x</sub> using Mathematica, Version 3.9 under Window98. I think that it is an interesting cuestion because many statistical packages give the asymptotic significance level in the statistical tests and with these porcedures one can easily calculate the exact significance levels and the left-tail and right-tail probabilities with non-parametric distributions.I have used mathematica to make these calculations because one can use symbolic language to solve recursion relations. It's very easy to generate the format of the tables, and it's possible to obtain any table of the mentioned non-parametric distributions with any precision, not only with the standard parameters more used in Statistics, and without transcription mistakes. Furthermore, using similar procedures, we can generate tables for the following distribution functions: Binomial, Poisson, Hypergeometric, Normal, x<sup>2</sup> Chi-Square, T-Student, F-Snedecor, Geometric, Gamma and Beta.
ER -