Jouvin L, Lemière A, Terrier R, Ohm S, Oya I, van Eldik C (2015)
Publication Language: English
Publication Status: Published
Publication Type: Conference contribution, Conference Contribution
Publication year: 2015
Publisher: Proceedings of Science (PoS)
Pages Range: 871
Article Number: 871
Conference Proceedings Title: 34th International Cosmic Ray Conference (ICRC2015)
URI: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84988709166&origin=inward
Spectral extraction in data from ground-based gamma-ray observations is usually performed using the on-off likelihood statistic which is based on the profile likelihood technique. The latter is known to lead to inconsistent estimators in some situations. We present here a systematic Monte Carlo study of the distribution of fitted spectral parameters for typical observations with VHE (very high energy; >100GeV) observatories and show that in some conditions spectral extraction yields inconsistent estimators. We discuss some techniques to alleviate this effect and the impact on the search for spectral cutoffs in the very low statistic regime.
APA:
Jouvin, L., Lemière, A., Terrier, R., Ohm, S., Oya, I., & van Eldik, C. (2015). Statistical biases of spectral analysis with the ON-OFF likelihood statistic. In 34th International Cosmic Ray Conference (ICRC2015) (pp. 871). Proceedings of Science (PoS).
MLA:
Jouvin, L., et al. "Statistical biases of spectral analysis with the ON-OFF likelihood statistic." Proceedings of the 34th International Cosmic Ray Conference, ICRC 2015 Proceedings of Science (PoS), 2015. 871.
BibTeX: Download