epäparametrisille

Epäparametriset refers to a category of statistical methods that do not rely on assumptions about the underlying distribution of the data. Unlike parametric methods, which often assume that data follows a specific distribution like the normal distribution, non-parametric methods are more flexible. They can be applied to a wider range of data types, including ordinal and nominal data, and are less sensitive to outliers.

The term "non-parametric" can be somewhat misleading, as some non-parametric methods do involve parameters, but these parameters are not tied to a specific distribution. Instead, they might relate to ranks or other characteristics of the data. These methods are sometimes referred to as distribution-free methods because their validity does not depend on the data conforming to a particular probability distribution.

Common examples of non-parametric tests include the Wilcoxon rank-sum test, the Mann-Whitney U test, and the Kruskal-Wallis test, which are often used as alternatives to the parametric t-test and ANOVA when their assumptions are violated. Other non-parametric techniques include kernel density estimation and bootstrapping. Their strength lies in their robustness and applicability when parametric assumptions are questionable or unknown. However, they can sometimes be less powerful than parametric tests when the assumptions of the parametric test are actually met.