Repository used to run the experiments described in the paper: Online Extra Trees Regressor

Mastelini, S. M., Nakano, F. K., Vens, C., & de Leon Ferreira, A. C. P. (2022). Online Extra Trees Regressor. IEEE Transactions on Neural Networks and Learning Systems.

Requirements

river>=0.14.0
river-extra>=0.14.0
pandas
numpy
matplotlib
seaborn

Repository organization

• datasets: evaluated datasets (data generators were also employed).
• output: folder with all the raw logs generated by the experiments.
• src: the scripts per se.

Scripts

The main scripts are run.py (datasets stored in files), run_drift.py (generator-based data), and saturation_study.py. The file baselines.py could also be interesing to check. All the parameters of the experiments are centralized in the the utils.py. The remaining scripts are used for generating plots, tables, and parsing the generated outputs.