Copula for Instance-wise Feature Selection and Rank

Hanyu Peng, Guanhua Fang, Ping Li
Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence, PMLR 216:1651-1661, 2023.

Abstract

Instance-wise feature selection and ranking methods can achieve a good selection of task-friendly features for each sample in the context of neural networks. However, existing approaches that assume feature subsets to be independent are imperfect when considering the dependency between features. To address this limitation, we propose to incorporate the Gaussian copula, a powerful mathematical technique for capturing correlations between variables, into the current feature selection framework with no additional changes needed. Experimental results on both synthetic and real datasets, in terms of performance comparison and interpretability, demonstrate that our method is capable of capturing meaningful correlations.

Cite this Paper

BibTeX
@InProceedings{pmlr-v216-peng23a, title = {Copula for Instance-wise Feature Selection and Rank}, author = {Peng, Hanyu and Fang, Guanhua and Li, Ping}, booktitle = {Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence}, pages = {1651--1661}, year = {2023}, editor = {Evans, Robin J. and Shpitser, Ilya}, volume = {216}, series = {Proceedings of Machine Learning Research}, month = {31 Jul--04 Aug}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v216/peng23a/peng23a.pdf}, url = {https://proceedings.mlr.press/v216/peng23a.html}, abstract = {Instance-wise feature selection and ranking methods can achieve a good selection of task-friendly features for each sample in the context of neural networks. However, existing approaches that assume feature subsets to be independent are imperfect when considering the dependency between features. To address this limitation, we propose to incorporate the Gaussian copula, a powerful mathematical technique for capturing correlations between variables, into the current feature selection framework with no additional changes needed. Experimental results on both synthetic and real datasets, in terms of performance comparison and interpretability, demonstrate that our method is capable of capturing meaningful correlations.} }
Endnote
%0 Conference Paper %T Copula for Instance-wise Feature Selection and Rank %A Hanyu Peng %A Guanhua Fang %A Ping Li %B Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2023 %E Robin J. Evans %E Ilya Shpitser %F pmlr-v216-peng23a %I PMLR %P 1651--1661 %U https://proceedings.mlr.press/v216/peng23a.html %V 216 %X Instance-wise feature selection and ranking methods can achieve a good selection of task-friendly features for each sample in the context of neural networks. However, existing approaches that assume feature subsets to be independent are imperfect when considering the dependency between features. To address this limitation, we propose to incorporate the Gaussian copula, a powerful mathematical technique for capturing correlations between variables, into the current feature selection framework with no additional changes needed. Experimental results on both synthetic and real datasets, in terms of performance comparison and interpretability, demonstrate that our method is capable of capturing meaningful correlations.
APA
Peng, H., Fang, G. & Li, P.. (2023). Copula for Instance-wise Feature Selection and Rank. Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 216:1651-1661 Available from https://proceedings.mlr.press/v216/peng23a.html.

Related Material