Abstract
In this paper, we study the performance of the Birnbaum–Saunders-power-exponential (BS-PE) kernel and Bayesian local bandwidth selection in the context of kernel density estimation for nonnegative heavy tailed data. Our approach considers the BS-PE kernel estimator and treats locally the bandwidth h as a parameter with prior distribution. The posterior density of h at each point x (point where the density is estimated) is derived in closed form, and the Bayesian bandwidth selector is obtained by using popular loss functions. The performance evaluation of this new procedure is carried out by a simulation study and real data in web-traffic. The proposed method is very quick and very competitive in comparison with the existing global methods, namely biased cross-validation and unbiased cross-validation.
Similar content being viewed by others
References
Birnbaum ZW, Saunders SC (1969) A new family of life distributions. J Appl Prob 6:319–327
Brewer MJ (2000) A Bayesian model for local smoothing in kernel density estimation. Stat Comput 10:299–309
Calabria R, Pulcin G (1996) Point estimation under asymmetric loss functions for left-truncated exponential samples. Commun Stat Theory Methods 25(3):585–600
Chen SX (1999) Beta kernel estimators for density functions. Comput Stat Data Anal 31:131–145
Chen SX (2000) Gamma kernel estimators for density functions. Ann Inst Stat Math 52:471–480
de Lima MS, Atuncar GS (2010) A Bayesian method to estimate the optimal bandwidth for multivariate kernel estimator. J Nonparametric Stat 23:137–148
Díaz-García JA, Leiva V (2005) A new family of life distributions based on the contoured elliptically distributions. J Stat Plan Inference 128:445–457
Fang KT, Kotz S, Ng WK (1990) Symmetric multivariate and related distributions. Chapman and Hall, London
Filippone M, Sanguinetti G (2011) Approximate inference of the bandwidth in multivariate kernel density estimation. Comput Stat Data Anal 55:3104–3122
Gangopadhyay AK, Cheung KN (2002) Bayesian approach to the choice of smoothing parameter in kernel density estimation. J Nonparametric Stat 14:655–664
Hu S, Poskitt DS, Zhang X (2012) Bayesian adaptive bandwidth kernel density estimation of irregular multivariate distributions. Comput Stat Data Anal 56:732–740
Igarashi G, Kakizawa Y (2014) Re-formulation of inverse Gaussian, reciprocal inverse Gaussian, and Birnbaum–Saunders kernel estimators. Stat Prob Lett 84:235–246
Jin X, Kawczak J (2003) Birnbaum–Saunders and lognormal kernel estimators for modelling durations in high frequency financial data. Ann Econ Finance 4:103–124
Kokonendji CC, Senga Kiessé T (2011) Discrete associated kernels method and extensions. Stat Methodol 8:497–516
Kulasekera KB, Padgett WJ (2006) Bayes bandwidth selection in kernel density estimation with censored data. J Nonparametric Stat 18:129–143
Kuruwita CN, Kulasekera KB, Padgett WJ (2010) Density estimation using asymmetric kernels and Bayes bandwidths with censored data. J Stat Plan Inference 140:1765–1774
Luz Gámiz M, Kulasekera KB, Limnios Nikolaos, Lindqvist Bo Henry (2011) Applied nonparametric statistics in reliability. Springer, London
Marchant C, Bertin K, Leiva V, Saulo H (2013) Generalized Birnbaum–Saunders kernel density estimators and an analysis of financial data. Comput Stat Data Anal 63:1–15
Markovitch NM, Krieger UR (2002) The estimation of heavy-tailed probability density functions, their mixtures and quantiles. Comput Netw 40:459–474
Parzen E (1962) On estimation of a probability density function and mode. Ann Math Stat 33:1065–1076
Robert CP (1996) Intrinsic losses. Theory Decis 40:191–214
Rosenblatt M (1956) Remarks on some nonparametric estimates of a density function. Ann Math Stat 27:832–837
Saulo H, Leiva V, Ziegelmann FA, Marchant C (2013) A nonparametric method for estimating asymmetric densities based on skewed Birnbaum–Saunders distributions applied to environmental data. Stoch Environ Res Risk Assess 7:1479–1491
Scaillet O (2004) Density estimation using inverse and reciprocal inverse Gaussian kernels. J Nonparametric Stat 16:217–226
Silverman BW (1986) Density estimation for statistics and data analysis. Chapman and Hall, New York
Simonoff JS (1996) Smoothing methods in statistics. Springer, New York
Wand M, Jones M (1995) Kernel smoothing. Chapman and Hall, London
Ziane Y, Adjabi S, Zougab N (2015) Adaptive Bayesian bandwidth selection in asymmetric kernel density estimation for nonnegative heavy-tailed data. J Appl Stat 42:8:1645–1658
Zhang X, King ML, Hyndman RJ (2006) A Bayesian approach to bandwidth selection for multivariate kernel density estimation. Comput Stat Data Anal 50:3009–3031
Zougab N, Adjabi S, Kokonendji CC (2012) Binomial kernel and Bayes local bandwidth in discrete functions estimation. J Nonparametric Stat 24:783–795
Zougab N, Adjabi S, Kokonendji CC (2013) Adaptive smoothing in associated kernel discrete functions estimation using Bayesian approach. J Stat Comput Simul 83:2219–2231
Zougab N, Adjabi S, Kokonendji CC (2014) Bayesian estimation of adaptive bandwidth matrices in multivariate kernel density estimation. Comput Stat Data Anal 75:28–38
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 Proof of Proposition 1
We here present the calculations of \(\hat{\pi }(h|x,X_{1},X_{2},\ldots ,X_{n})\) and \(\hat{h}_{n}(x)\) (with quadratic loss function) using the prior (13) and the BS-PE kernel density estimator (5):
Now, we consider the denominator of (18),
Let \(u=\frac{1}{h^{\nu }}\). Then we can rewrite the above integral as:
The local bandwidth \(\hat{h}_{n}(x)\) is given by
By the same reasoning for the numerator of (21), the local bandwidth is then given by:
which corresponds to Eq. (15). Similarly we can use the same technique above for the case of \(\tilde{h}_{n}(x)\) given by (16) which completes the proof.
Rights and permissions
About this article
Cite this article
Ziane, Y., Zougab, N. & Adjabi, S. Birnbaum–Saunders power-exponential kernel density estimation and Bayes local bandwidth selection for nonnegative heavy tailed data. Comput Stat 33, 299–318 (2018). https://doi.org/10.1007/s00180-017-0712-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-017-0712-8