On the Chance Accuracies of Large Collections of Classifiers
Abstract
We provide a theoretical analysis of the chance accuracies of large collections of classifiers. We show that on problems with small numbers of examples, some classifier can perform well by random chance, and we derive a theorem to explicitly calculate this accuracy. We use this theorem to provide a principled feature selection criteria for sparse, high-dimensional problems. We evaluate this method on both microarray and fMRI datasets and show that it performs very close to the optimal accuracy obtained from an oracle. We also show that on the fMRI dataset this technique chooses relevant features successfully while another state-of-the-art method, the False Discovery Rate (FDR), completely fails at standard significance levels.
Cite
Text
Palatucci and Carlson. "On the Chance Accuracies of Large Collections of Classifiers." International Conference on Machine Learning, 2008. doi:10.1145/1390156.1390250Markdown
[Palatucci and Carlson. "On the Chance Accuracies of Large Collections of Classifiers." International Conference on Machine Learning, 2008.](https://mlanthology.org/icml/2008/palatucci2008icml-chance/) doi:10.1145/1390156.1390250BibTeX
@inproceedings{palatucci2008icml-chance,
title = {{On the Chance Accuracies of Large Collections of Classifiers}},
author = {Palatucci, Mark and Carlson, Andrew},
booktitle = {International Conference on Machine Learning},
year = {2008},
pages = {744-751},
doi = {10.1145/1390156.1390250},
url = {https://mlanthology.org/icml/2008/palatucci2008icml-chance/}
}