Nonparametric Statistics
A better name for this topic is distribution-free Statistics, as we would like to call it. Here are some of the R tutorials in markdown format that I have written for this class.
Most of these examples were collected from various sources / books etc. and I have tried to give credits for each of them. However, if you notice something with a missing citation or you think should not be here, please let me know at jd033@uark.edu.
This website is for class purposes only.
Syllabus
- Motivation for Nonparametric methods.
- Review of univariate probability & central limit theorem.
- Review of parametric inference: Point & interval estimation and Hypothesis testing (P-value).
- Tests of randomness**
- Tests of goodness of fit (Chi square goodness-of-fit tests. Kolmogorov-Smirnov goodness-of-fit tests, Lilliefors test for Normality).
- One-sample and paired-sample procedures (Sign test & the Wilcoxon signed-rank test, normal approximation and continuity correction, calculation of exact P-values).
- Two-sample problem: Location model, scale model, Permutation test, The Wilcoxon rank-sum test /Mann-Whitney U test, Discuss K-S test for two samples, Calculation of exact P-value and normal approximation. (Equivalence of rank-sum & Mann-Whitney U test). Distribution-free property of Kolmogorov-Smirnov Test Statistics and Asymptotic distribution (Brownian Bridge).
- Test for scale model: Siegel-Tukey & similar tests* Additional topics: Linear rank tests, van der Waerden test, Asymptotic Relative Efficiency, Randomized Test.
- Tests of equality of K distributions (Kruskal-Wallis one-way ANOVA)
- Various nonparametric measures of association (Kendall’s \(\tau\), Spearman’s \(\rho\)).
- Measures of association in multiple classifications.
- Bootstrap & other resampling techniques*
- Analysis of count data*