A Robust-Equitable Copula Dependence Measure for Feature Selection

Yale Chang, Yi Li, Adam Ding, Jennifer Dy
Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, PMLR 51:84-92, 2016.

Abstract

Feature selection aims to select relevant features to improve the performance of predictors. Many feature selection methods depend on the choice of dependence measures. To select features that have complex nonlinear relationships with the response variable, the dependence measure should be equitable: treating linear and nonlinear relationships equally. In this paper we introduce the concept of robust-equitability and a robust-equitable dependence measure copula correlation (Ccor). This measure has the following advantages compared to existing dependence measures: it is robust to different relationship forms and robust to unequal sample sizes of different features. In contrast, existing dependence measures cannot take these factors into account simultaneously. Experiments on synthetic and real-world datasets confirm our theoretical analysis, and illustrates its advantage in feature selection.

Cite this Paper


BibTeX
@InProceedings{pmlr-v51-chang16, title = {A Robust-Equitable Copula Dependence Measure for Feature Selection}, author = {Chang, Yale and Li, Yi and Ding, Adam and Dy, Jennifer}, booktitle = {Proceedings of the 19th International Conference on Artificial Intelligence and Statistics}, pages = {84--92}, year = {2016}, editor = {Gretton, Arthur and Robert, Christian C.}, volume = {51}, series = {Proceedings of Machine Learning Research}, address = {Cadiz, Spain}, month = {09--11 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v51/chang16.pdf}, url = {https://proceedings.mlr.press/v51/chang16.html}, abstract = {Feature selection aims to select relevant features to improve the performance of predictors. Many feature selection methods depend on the choice of dependence measures. To select features that have complex nonlinear relationships with the response variable, the dependence measure should be equitable: treating linear and nonlinear relationships equally. In this paper we introduce the concept of robust-equitability and a robust-equitable dependence measure copula correlation (Ccor). This measure has the following advantages compared to existing dependence measures: it is robust to different relationship forms and robust to unequal sample sizes of different features. In contrast, existing dependence measures cannot take these factors into account simultaneously. Experiments on synthetic and real-world datasets confirm our theoretical analysis, and illustrates its advantage in feature selection.} }
Endnote
%0 Conference Paper %T A Robust-Equitable Copula Dependence Measure for Feature Selection %A Yale Chang %A Yi Li %A Adam Ding %A Jennifer Dy %B Proceedings of the 19th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2016 %E Arthur Gretton %E Christian C. Robert %F pmlr-v51-chang16 %I PMLR %P 84--92 %U https://proceedings.mlr.press/v51/chang16.html %V 51 %X Feature selection aims to select relevant features to improve the performance of predictors. Many feature selection methods depend on the choice of dependence measures. To select features that have complex nonlinear relationships with the response variable, the dependence measure should be equitable: treating linear and nonlinear relationships equally. In this paper we introduce the concept of robust-equitability and a robust-equitable dependence measure copula correlation (Ccor). This measure has the following advantages compared to existing dependence measures: it is robust to different relationship forms and robust to unequal sample sizes of different features. In contrast, existing dependence measures cannot take these factors into account simultaneously. Experiments on synthetic and real-world datasets confirm our theoretical analysis, and illustrates its advantage in feature selection.
RIS
TY - CPAPER TI - A Robust-Equitable Copula Dependence Measure for Feature Selection AU - Yale Chang AU - Yi Li AU - Adam Ding AU - Jennifer Dy BT - Proceedings of the 19th International Conference on Artificial Intelligence and Statistics DA - 2016/05/02 ED - Arthur Gretton ED - Christian C. Robert ID - pmlr-v51-chang16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 51 SP - 84 EP - 92 L1 - http://proceedings.mlr.press/v51/chang16.pdf UR - https://proceedings.mlr.press/v51/chang16.html AB - Feature selection aims to select relevant features to improve the performance of predictors. Many feature selection methods depend on the choice of dependence measures. To select features that have complex nonlinear relationships with the response variable, the dependence measure should be equitable: treating linear and nonlinear relationships equally. In this paper we introduce the concept of robust-equitability and a robust-equitable dependence measure copula correlation (Ccor). This measure has the following advantages compared to existing dependence measures: it is robust to different relationship forms and robust to unequal sample sizes of different features. In contrast, existing dependence measures cannot take these factors into account simultaneously. Experiments on synthetic and real-world datasets confirm our theoretical analysis, and illustrates its advantage in feature selection. ER -
APA
Chang, Y., Li, Y., Ding, A. & Dy, J.. (2016). A Robust-Equitable Copula Dependence Measure for Feature Selection. Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 51:84-92 Available from https://proceedings.mlr.press/v51/chang16.html.

Related Material