03 Dec Introduction to non-parametric methods
1:00 pm - 2:30 pm
statistics seminar | level: intermediate |
for questions related to this event, contact email@example.com
affiliation: Ghent University
This seminar is about the essentials of non-parametric methods. Non-parametric tests are more suitable for data that come from skewed distributions or have a discrete or ordinal scale. They are also the method of choice for small sample sized data.
The first section is about non-parametric hypothesis testing wherein methods such as the sign test (one-sample), Wilcoxon Rank Sum Test (for 2 samples) and the Kruskal–Wallis Test (more than 2 samples) and presented. The aforementioned tests for 2 or more sample all assume that the observations are independently sampled. For dependent samples, alternatives such as the Wilcoxon Signed Rank Test ( for 2 samples) and the Friedman test (for more than 2 samples) should be considered.
The second section is about non-parametric measures for gauging the strength of associations between 2 variables with special attention to the situation where the two variables can be all continuous, one continuous and one categorical and the case where both are categorical.
Finally, section 3 is about non-parametric regression where the linearity assumption which characterizes linear models is relaxed. Different types of non-parametric regression models will be explained with demonstrations of how to fit and assess the models.
The examples will be demonstrated using RStudio and all codes will be provided