A cluster based hybrid feature selection approach

Pablo A. Jaskowiak, Ricardo J.G.B. Campello

Publikation: Kapitel i bog/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review


Data collection and storage capacities have increased significantly in the past decades. In order to cope with the increasingly complexity of data, feature selection methods have become an omnipresent preprocessing step in data analysis. In this paper we present a hybrid (filter - wrapper) feature selection method tailored for data classification problems. Our hybrid approach is composed of two stages. In the first stage, a filter clusters features to identify and remove redundancy. In the second stage, a wrapper evaluates different feature subsets produced by the filter, determining the one that produces the best classification performance in terms of accuracy. The effectiveness of our method is demonstrated through an empirical evaluation performed on real-world datasets coming from various sources.

TitelProceedings - 2015 Brazilian Conference on Intelligent Systems, BRACIS 2015
Publikationsdato2. mar. 2016
ISBN (Elektronisk)9781509000166
StatusUdgivet - 2. mar. 2016
Udgivet eksterntJa
Begivenhed4th Brazilian Conference on Intelligent Systems, BRACIS 2015 - Natal, Brasilien
Varighed: 4. nov. 20157. nov. 2015


Konference4th Brazilian Conference on Intelligent Systems, BRACIS 2015
SponsorBrazilian Computer Society (SBC)

Bibliografisk note

Funding Information:
This project was partially funded by Brazilian Research Agencies FAPESP and CNPq. Pablo A. Jaskowiak thanks FAPESP (Grant #2011/04247-5). Ricardo J. G. B. Campello thanks FAPESP (Grant #2013/18698-4) and CNPq (Grant #304137/2013-8).

Publisher Copyright:
© 2015 IEEE.


Dyk ned i forskningsemnerne om 'A cluster based hybrid feature selection approach'. Sammen danner de et unikt fingeraftryk.