Advances in Data Mining Knowledge Discovery and Applications

0380642 - ÚI 2013 RIV HR eng M - Část monografie knihy
Borovička, T. - Jiřina jr., M. - Kordík, P. - Jiřina, Marcel
Selecting Representative Data Sets.
Advances in Data Mining Knowledge Discovery and Applications. Rijeka: InTech, 2012 - (Karahoca, A.), s. 43-70. ISBN 978-953-51-0748-4
Grant CEP: GA MŠMT(CZ) LG12020
Institucionální podpora: RVO:67985807
Klíčová slova: data selection * classification * class balancing * sampling
Kód oboru RIV: BB - Aplikovaná statistika, operační výzkum

The aim of the chapter is to give an exhaustive overview and comparision of existing methods that deal with the methods of data selection and sampling. A general approach to the problem of optimal data selection (we could call it also splitting, dividing, sampling, ...) to training, testing and eventually validation sets is discussed. An overview of the methods together with their features, utilization, positives and negatives is given. Aptly presented algorithms of the methods are clearly summarized. Principles of selected methods are visualized in pictures and charts.
Trvalý link: http://hdl.handle.net/11104/0211296