Publication detail

Comparison of Wavelet Based Feature Extraction Methods in Classification of Emission Line Stellar Spectra

VRÁBELOVÁ, P.

Original Title

Comparison of Wavelet Based Feature Extraction Methods in Classification of Emission Line Stellar Spectra

English Title

Comparison of Wavelet Based Feature Extraction Methods in Classification of Emission Line Stellar Spectra

Type

conference paper

Language

en

Original Abstract

The paper compares different wavelet transforms, wavelets, and statistical metrics in attempt to find the best feature extraction method for classification of Be star spectra using support vector machines (SVM).

English abstract

The paper compares different wavelet transforms, wavelets, and statistical metrics in attempt to find the best feature extraction method for classification of Be star spectra using support vector machines (SVM).

Keywords

Be star, stellar spectrum, emission line, feature extraction, wavelet transform, classification, support vector machines, SVM

RIV year

2014

Released

04.03.2014

Publisher

Brno University of Technology

Location

Brno

ISBN

978-80-214-4924-4

Book

Proceedings of the 20th Conference STUDENT EEICT 2014

Edition

Volume 3

Edition number

NEUVEDEN

Pages from

254

Pages to

258

Pages count

5

URL

Documents

BibTex


@inproceedings{BUT111539,
  author="Pavla {Vrábelová}",
  title="Comparison of Wavelet Based Feature Extraction Methods in Classification of Emission Line Stellar Spectra",
  annote="The paper compares different wavelet transforms, wavelets, and statistical
metrics in attempt to find the best feature extraction method for classification
of Be star spectra using support vector machines (SVM).",
  address="Brno University of Technology",
  booktitle="Proceedings of the 20th Conference STUDENT EEICT 2014",
  chapter="111539",
  edition="Volume 3",
  howpublished="print",
  institution="Brno University of Technology",
  year="2014",
  month="march",
  pages="254--258",
  publisher="Brno University of Technology",
  type="conference paper"
}