Konferenční příspěvek

Somol P., Novovičová J., Pudil P.

Proc. 6th IASTED Int. Conf. on Advances in Computer Science and Engineering

ACTA Press

Calgary

2010

978-0-88986-830-4

Doc. RNDr. Jana Novovičová, CSc. (ÚTIA - Oddělení rozpoznávání obrazů)

Feature selection, sequential search, hybrid methods, classi

In this paper we propose the general scheme of defining hybrid feature selection algorithms based on standard sequential search with the aim to improve feature selection performance, especially on high-dimensional or large-sample data. We show experimentally that “hybridization” has not only the potential to dramatically reduce FS search time, but in some cases also to actually improve classifier generalization, i.e., its classification performance on previously unknown data.