Closed-Form Dual Perturb and Combine for Tree-Based Models
Abstract
This paper studies the aggregation of predictions made by tree-based models for several perturbed versions of the attribute vector of a test case. A closed-form approximation of this scheme combined with cross-validation to tune the level of perturbation is proposed. This yields soft-tree models in a parameter free way. and preserves their interpretability. Empirical evaluations, on classification and regression problems, show that accuracy and bias/variance tradeoff are improved significantly at the price of an acceptable computational overhead. The method is further compared and combined with tree bagging.
Cite
Text
Geurts and Wehenkel. "Closed-Form Dual Perturb and Combine for Tree-Based Models." International Conference on Machine Learning, 2005. doi:10.1145/1102351.1102381Markdown
[Geurts and Wehenkel. "Closed-Form Dual Perturb and Combine for Tree-Based Models." International Conference on Machine Learning, 2005.](https://mlanthology.org/icml/2005/geurts2005icml-closed/) doi:10.1145/1102351.1102381BibTeX
@inproceedings{geurts2005icml-closed,
title = {{Closed-Form Dual Perturb and Combine for Tree-Based Models}},
author = {Geurts, Pierre and Wehenkel, Louis},
booktitle = {International Conference on Machine Learning},
year = {2005},
pages = {233-240},
doi = {10.1145/1102351.1102381},
url = {https://mlanthology.org/icml/2005/geurts2005icml-closed/}
}