Supervised Feature Selection via Dependence Estimation - Computer Science LearningReportar como inadecuado




Supervised Feature Selection via Dependence Estimation - Computer Science Learning - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: We introduce a framework for filtering features that employs theHilbert-Schmidt Independence Criterion HSIC as a measure of dependencebetween the features and the labels. The key idea is that good features shouldmaximise such dependence. Feature selection for various supervised learningproblems including classification and regression is unified under thisframework, and the solutions can be approximated using a backward-eliminationalgorithm. We demonstrate the usefulness of our method on both artificial andreal world datasets.



Autor: Le Song, Alex Smola, Arthur Gretton, Karsten Borgwardt, Justin Bedo

Fuente: https://arxiv.org/







Documentos relacionados