@inproceedings{663f51dcbde44f8ead1ae96a6a675c53,
title = "Evaluating feature selection for SVMs in high dimensions",
abstract = "We perform a systematic evaluation of feature selection (FS) methods for support vector machines (SVMs) using simulated high-dimensional data (up to 5000 dimensions). Several findings previously reported at low dimensions do not apply in high dimensions. For example, none of the FS methods investigated improved SVM accuracy, indicating that the SVM built-in regularization is sufficient. These results were also validated using microarray data. Moreover, all FS methods tend to discard many relevant features. This is a problem for applications such as microarray data analysis, where identifying all biologically important features is a major objective.",
author = "Roland Nilsson and Pe{\~n}a, {Jos{\'e} M.} and Johan Bj{\"o}rkegren and Jesper Tegn{\'e}r",
year = "2006",
doi = "10.1007/11871842_72",
language = "English (US)",
isbn = "354045375X",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "719--726",
booktitle = "Machine Learning",
address = "Germany",
note = "17th European Conference on Machine Learning, ECML 2006 ; Conference date: 18-09-2006 Through 22-09-2006",
}