Improving Algorithm-Selectors and Performance-Predictors via Learning Discriminating Training Samples - Code and Data

This repository contains the code and data for reproducibility of the paper 'Improving Algorithm-Selectors and Performance-Predictors via Learning Discriminating Training Samples'. The following files are included: Plots: Additional plots not in the paper; Code: Python scripts to generate trajectories and perform classification/regression; best_algo.csv : Labels for the classification; performances.csv : Performances used for the regression; SA_parameters.csv : SA parameters for all machine learning tasks; irace_scenario.txt : scenario used for the tuning.