Eva García-Martín(1), Niklas Lavesson(1,2), Håkan Grahn(1), Emiliano Casalicchio(1,3), and Veselka Boeva(1)
(1) Blekinge Institute of Technology, Karlskrona, Sweden
(2) Jönköping University, Jönköping, Sweden
(3) Sapienza University of Rome, Rome, Italy
Paper accepted at 2018 IEEE International Conference on Data Science and Advanced Analytics (DSAA) (to appear)
Extension of this paper titled "Energy Aadaptive Very Fast Decision Tree" submitted to "International Journal of Data Science and Analytics (JDSA)", experiments in folder: code/experiments/JDSA_18
Code: code/VFDT-nmin/original/src/learners/avfdt
Experiments:
- DSAA: code/experiments/DSAA18/{run.sh,create_plots.ipynb,create_tables.ipynb}
- JDSA: code/experiments/JDSA_18/
VFDT, CVFDT, VFDT-nmin comparison
15 datasets
Run: 10 times and averaged the results
Energy measurement: Intel Power Gadget https://software.intel.com/en-us/articles/intel-power-gadget-20
Script that runs the experiments: code/experiments/DSAA18/run.sh
VFDT, CVFDT, VFDT-nmin comparison
29 datasets
Run: 5 times and averaged the results
Energy measurement: Intel Power Gadget https://software.intel.com/en-us/articles/intel-power-gadget-20
Script that runs the experiments: code/experiments/JDSA_18/{run_baseline.sh,run_drift.sh,run_real.sh}
@inproceedings{garcia2018hoeffding,
title={Hoeffding Trees with nmin adaptation},
author={Garcia-Martin, Eva and Lavesson, Niklas and Grahn, H{\aa}kan and Casalicchio, Emiliano and Boeva,Veselka},
booktitle={Data Science and Advanced Analytics (DSAA), 2018 IEEE International Conference on},
year={2018},
organization={IEEE}
}