Globally Induced Forest: A Prepruning Compression Scheme

Jean-Michel Begon, Arnaud Joly, Pierre Geurts
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:420-428, 2017.

Abstract

Tree-based ensemble models are heavy memory-wise. An undesired state of affairs considering nowadays datasets, memory-constrained environment and fitting/prediction times. In this paper, we propose the Globally Induced Forest (GIF) to remedy this problem. GIF is a fast prepruning approach to build lightweight ensembles by iteratively deepening the current forest. It mixes local and global optimizations to produce accurate predictions under memory constraints in reasonable time. We show that the proposed method is more than competitive with standard tree-based ensembles under corresponding constraints, and can sometimes even surpass much larger models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-begon17a, title = {Globally Induced Forest: A Prepruning Compression Scheme}, author = {Jean-Michel Begon and Arnaud Joly and Pierre Geurts}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {420--428}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/begon17a/begon17a.pdf}, url = {https://proceedings.mlr.press/v70/begon17a.html}, abstract = {Tree-based ensemble models are heavy memory-wise. An undesired state of affairs considering nowadays datasets, memory-constrained environment and fitting/prediction times. In this paper, we propose the Globally Induced Forest (GIF) to remedy this problem. GIF is a fast prepruning approach to build lightweight ensembles by iteratively deepening the current forest. It mixes local and global optimizations to produce accurate predictions under memory constraints in reasonable time. We show that the proposed method is more than competitive with standard tree-based ensembles under corresponding constraints, and can sometimes even surpass much larger models.} }
Endnote
%0 Conference Paper %T Globally Induced Forest: A Prepruning Compression Scheme %A Jean-Michel Begon %A Arnaud Joly %A Pierre Geurts %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-begon17a %I PMLR %P 420--428 %U https://proceedings.mlr.press/v70/begon17a.html %V 70 %X Tree-based ensemble models are heavy memory-wise. An undesired state of affairs considering nowadays datasets, memory-constrained environment and fitting/prediction times. In this paper, we propose the Globally Induced Forest (GIF) to remedy this problem. GIF is a fast prepruning approach to build lightweight ensembles by iteratively deepening the current forest. It mixes local and global optimizations to produce accurate predictions under memory constraints in reasonable time. We show that the proposed method is more than competitive with standard tree-based ensembles under corresponding constraints, and can sometimes even surpass much larger models.
APA
Begon, J., Joly, A. & Geurts, P.. (2017). Globally Induced Forest: A Prepruning Compression Scheme. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:420-428 Available from https://proceedings.mlr.press/v70/begon17a.html.

Related Material