A Study on the Use of Non-Parametric Tests for Experimentation with Cluster Analysis

In this paper, we focus on the experimental analysis on the performance in cluster analysis with the use of nonparametric tests on the clustering task. Particularly, we have studied whether the sample of results from multiple trials obtained by conventional clustering algorithms checks the necessary conditions for being analyzed through parametrical tests. The study is conducted by considering the possibilities on clustering experiments. The study obtained state that a parametric statistical analysis could not be appropriate specially when we deal with multiple-problem results. In multiple-problem analysis, we propose the use of non parametric statistical tests given that they are less restrictive than parametric ones and they can be used over small size samples of results. These conditions are problem-dependent and indefinite, which justifies the need of using nonparametric statistics in the experimental analysis.