This repository contains the generator python code for the artificial datasets and some results of the Hostility measure from paper "Hostility measure for multi-level study of data complexity".
- Artificial_data has the code to generate the bivariate and multivariate Normal artifical data.
- Algorithm_code contains a script with the code of the Hostility measure and an example using artificial data from previous folder for running it.
- Visual_analysis collects the graphical analysis comparing different state-of-art complexity measures for all artificial datasets.
- Filter_experiment contains, for all datasets, the evolution of errors when the more complex points are eliminated.