A.I. Krasilnikov
Èlektron. model. 2017, 39(4):19-30
https://doi.org/10.15407/emodel.39.04.019
ABSTRACT
A formula for finding the kurtosis coefficient of symmetric contaminated Gaussian distributions has been obtained. The dependence of the kurtosis coefficient on the parameters of the model of contaminated distributions has been studied. Examples of contaminating with uniform and logistic distributions have been considered. The obtained results allow the author to analyze non-Gaussian random variables described by the model of contaminated Gaussian distributions.
KEYWORDS
contaminated distributions, Tukey-Huber model, mixtures of distributions, kurtosis coefficient, cumulant analysis.
REFERENCES
1. Aivazian, S.A., Eniukov, I.S. and Meshalkin, L.D. (1983), Prikladnaya statistika: Osnovy modelirovaniia i pervichnaia obrabotka dannykh. Spravochnoe izd. [Applied statistics: bases of modeling and initial data processing. Reference edition], Finansy i statistika, Moscow, USSR.
2. Mukha, V.S. (2009), Statisticheskie metody obrabotki dannykh: Uchebnoe posobie [Statistical methods of data processing: Tutorial], Izdatelskiy tsentr BGU, Minsk, Belarus.
3. Lemeshko, B.Yu., Lemeshko, S.B., Postovalov, S.N. and Chimitova, E.V. (2011), Statisticheskii analiz dannykh, modelirovanie i issledovanie veroiatnostnykh zakonomernostei. Kompiuternyi podkhod [Statistical data analysis, simulation and study of probability regularities. Computer approach], Izdatelstvo NGTU, Novosibirsk, Russia.
4. Tukey, J.W. (1960), A survey of sampling from contaminated distributions. Contributions to Probability and Statistics, Ed. I. Olkin, Stanford University Press, Stanford, UK.
5. Huber, P.J. (1984), Robastnost v statistike [Robust statistics], Translated by I.A. Makhova and V.I. Khokhlov, Ed. I.G. Zhurbenko, Mir, Moscow, USSR.
6. Hampel, F., Ronchetti, E., Rousseeuw, P. and Stahel, W. (1989), Robastnost v statistike. Podkhod na osnove funktsii vliianiia [Robust statistics. The approach based on influence functions], Translated by V.M. Zolotarev, Mir, Moscow, USSR.
7. Figueiredo, F. and Gomes, M.I. (2016), The total median statistic to monitor contaminated normal data, Journal Quality Technology & Quantitative Management, Vol. 13, pp. 1-16, available at: http://www.tandfonline.com/doi/abs/10.1080/16843703.2016.1139840
8. Punzo, A. and McNicholas, P.D. (2016), Parsimonious mixtures of multivariate contaminated normal distributions, Preprint submitted to arXiv 1305.4669, May 20, 2016. - pp. 1-28, available at: https://arxiv.org/pdf/1305.4669.pdf
9. Marchuk, V.I. and Tokareva, S.V. (2009), Sposoby obnaruzheniia anomalnykh znachenii pri analize nestatsionarnykh sluchainykh protsessov: Monografiia [Methods for detecting anomalous values in the analysis of non-stationary random processes: Monograph], Yuzhnorossiiskii gosudarstvennyi universitet ekonomiki i servisa, Shakhty, Russia.
10. Denisov, V.I. and Timofeev, V.S. (2011), “Stable distributions and estimation of parameters of regression dependencies”, Izvestiya Tomskogo politekhnicheskogo instituta, Vol. 318, no. 2, pp. 10-15.
11. Osadchiy, I.S. (2015), “Method for estimating the distribution parameters of Gaussian noise for the operation of a pulse signal system”, Zhurnal radioelektroniki: elektronnyy zhurnal, no. 4, pp. 1-27, available at: http://jre.cplire.ru/jre/apr15/1/text.html
12. Robbins, H. (1948), Mixture of distributions, Ann. Math. Stat, Vol. 19, pp. 360-369.
https://doi.org/10.1214/aoms/1177730200
13. Titterington, D.M., Smith, A.F.M. and Makov, U.E. (1985), Statistical analysis of finite mixture distributions, John Wiley & Sons, New York, USA.
14. McLachlan, G. and Peel, D. (2002), Finite mixture models, John Wiley & Sons, New York, USA.
15. Korolev, V.Yu. (2004), Smeshannye gaussovskie veroiatnostnye modeli realnykh protsessov [The mixed Gaussian probabilistic models of real processes], Maks Press, Moscow, Russia.
16. Korolev, V.Yu. (2008), Veroyatnostno-statisticheskii analiz khaoticheskikh protsessov s pomoshchiu smeshannykh gaussovskikh modelei. Dekompozitsiia volatilnosti finansovykh indeksov i turbulentnoi plazmy [Probabilistic-statistical analysis of chaotic processes using mixed Gaussian models. Decomposition of volatility of financial indices and turbulent plasma], Izdatelstvo Instituta problem informatiki RAN, Moscow, Russia.
17. Krasilnikov, A.I. and Pilipenko, K.P. (2007), “Unimodal two-componental Gaussian mixture. Excess kurtosis”, Elektronika i sviaz, no. 2 (37), pp. 32-38.
18. Krasilnikov, A.I. (2017), “Class of non-Gaussian symmetric distributions with zero coefficient of kurtosis”, Elektronnoe modelirovanie, Vol. 39, no. 1, pp. 3-17.
19. Vadzinskii, R.N. (2001), Spravochnik po veroiatnostnym raspredeleniiam [Reference book on probabilistic distributions], Nauka, St. Petersburg, Russia.