Approximate Bayesian Computation (ABC for short) is a family of computational techniques which offer an almost automated solution in situations where evaluation of the posterior likelihood is computationally prohibitive, or whenever suitable likelihoods are not available. In the present paper, we analyze the procedure from the point of view of -nearest neighbor theory and explore the statistical properties of its outputs. We discuss in particular some asymptotic features of the genuine conditional density estimate associated with ABC, which is an interesting hybrid between a -nearest neighbor and a kernel method.
Le terme anglais « Approximate Bayesian Computation » (ABC en abrégé) désigne une famille de techniques bayésiennes ayant pour objet la simulation selon une loi de probabilité lorsque la vraisemblance a posteriori n’est pas disponible ou s’avère impossible à évaluer numériquement. Dans le présent article, nous envisageons cette procédure du point de vue de la théorie des -plus proches voisins, en nous attachant plus particulièrement à examiner les propriétés statistiques des sorties de l’algorithme. Cela nous conduit à analyser le comportement asymptotique d’un estimateur de la densité conditionnelle naturellement associé à ABC, utilisé en pratique et possédant à la fois les caractéristiques d’un estimateur des -plus proches voisins et celles d’une méthode à noyau.
Keywords: approximate bayesian computation, nonparametric estimation, conditional density estimation, nearest neighbor methods, mathematical statistics
@article{AIHPB_2015__51_1_376_0, author = {Biau, G\'erard and C\'erou, Fr\'ed\'eric and Guyader, Arnaud}, title = {New insights into {Approximate} bayesian {Computation}}, journal = {Annales de l'I.H.P. Probabilit\'es et statistiques}, pages = {376--403}, publisher = {Gauthier-Villars}, volume = {51}, number = {1}, year = {2015}, doi = {10.1214/13-AIHP590}, zbl = {06412909}, language = {en}, url = {http://www.numdam.org/articles/10.1214/13-AIHP590/} }
TY - JOUR AU - Biau, Gérard AU - Cérou, Frédéric AU - Guyader, Arnaud TI - New insights into Approximate bayesian Computation JO - Annales de l'I.H.P. Probabilités et statistiques PY - 2015 SP - 376 EP - 403 VL - 51 IS - 1 PB - Gauthier-Villars UR - http://www.numdam.org/articles/10.1214/13-AIHP590/ DO - 10.1214/13-AIHP590 LA - en ID - AIHPB_2015__51_1_376_0 ER -
%0 Journal Article %A Biau, Gérard %A Cérou, Frédéric %A Guyader, Arnaud %T New insights into Approximate bayesian Computation %J Annales de l'I.H.P. Probabilités et statistiques %D 2015 %P 376-403 %V 51 %N 1 %I Gauthier-Villars %U http://www.numdam.org/articles/10.1214/13-AIHP590/ %R 10.1214/13-AIHP590 %G en %F AIHPB_2015__51_1_376_0
Biau, Gérard; Cérou, Frédéric; Guyader, Arnaud. New insights into Approximate bayesian Computation. Annales de l'I.H.P. Probabilités et statistiques, Volume 51 (2015) no. 1, pp. 376-403. doi : 10.1214/13-AIHP590. http://www.numdam.org/articles/10.1214/13-AIHP590/
[1] On bandwidth variation in kernel estimates – A square root law. Ann. Statist. 10 (1982) 1217–1223. | DOI | MR | Zbl
.[2] Bandwidth selection for kernel conditional density estimation. Comput. Statist. Data Anal. 36 (2001) 279–298. | DOI | MR | Zbl
and .[3] Adaptive approximate Bayesian computation. Biometrika 96 (2009) 983–990. | DOI | MR | Zbl
, , and .[4] Approximate Bayesian computation in population genetics. Genetics 162 (2002) 2025–2035. | DOI
, and .[5] On the rate of convergence of the bagged nearest neighbor estimate. J. Mach. Learn. Res. 11 (2010) 687–712. | MR | Zbl
, and .[6] Approximate Bayesian computation: A nonparametric perspective. J. Amer. Statist. Assoc. 105 (2010) 1178–1187. | DOI | MR | Zbl
.[7] Variable kernel estimates of multivariate densities. Technometrics 19 (1977) 135–144. | DOI | Zbl
, and .[8] Nearest neighbor classification in infinite dimension. ESAIM Probab. Stat. 10 (2006) 340–355. | DOI | Numdam | MR | Zbl
and .[9] Estimation by the nearest neighbor rule. IEEE Trans. Inform. Theory 14 (1968) 50–55. | DOI | Zbl
.[10] Differentiation of Integrals in . Lecture Notes in Mathematics 481. Springer, Berlin, 1975. | MR | Zbl
.[11] Necessary and sufficient conditions for the pointwise convergence of nearest neighbor regression function estimates. Z. Wahrsch. Verw. Gebiete 61 (1982) 467–481. | DOI | MR | Zbl
.[12] New multivariate product density estimates. J. Multivariate Anal. 82 (2002) 88–110. | DOI | MR | Zbl
and .[13] A Probabilistic Theory of Pattern Recognition. Springer, New York, 1996. | DOI | MR | Zbl
, and .[14] A crossvalidation method for estimating conditional densities. Biometrika 91 (2004) 819–834. | DOI | MR | Zbl
and .[15] A quantile-copula approach to conditional density estimation. J. Multivariate Anal. 100 (2009) 2083–2099. | DOI | MR | Zbl
.[16] Constructing summary statistics for approximate Bayesian computation: Semi-automatic approximate Bayesian computation. J. Roy. Statist. Soc. Ser. B 74 (2012) 419–474. | DOI | MR | Zbl
and .[17] Discriminatory analysis – Nonparametric discrimination: Consistency properties. Project 21-49-004, Report Number 4, USAF School of Aviation Medicine, Randolph Field, TX, 1951. | Zbl
and .[18] Estimating the age of the common ancestor of a sample of DNA sequences. Mol. Biol. Evol. 14 (1997) 195–199. | DOI
and .[19] Nonparametric estimation of conditional distributions. IEEE Trans. Inform. Theory 53 (2007) 1872–1879. | DOI | MR | Zbl
and .[20] Variable window width kernel estimates of probability densities. Probab. Theory Related Fields 80 (1988) 37–49. | DOI | MR | Zbl
and .[21] Cross-validation and the estimation of conditional probability densities. J. Amer. Statist. Assoc. 99 (2004) 1015–1026. | DOI | MR | Zbl
, and .[22] Nonparametric conditional density estimation. Technical report, Univ. Wisconsin, 2004.
.[23] Inequalities. Cambridge Univ. Press, Cambridge, 1988. | MR | Zbl
, and .[24] Monte Carlo sampling methods using Markov chains and their applications. Biometrika 57 (1970) 97–109. | DOI | MR | Zbl
.[25] Estimating and visualizing conditional densities. J. Comput. Graph. Statist. 5 (1996) 315–336. | MR
, and .[26] Note on the differentiability of multiple integrals. Fund. Math. 25 (1935) 217–234. | DOI | EuDML | JFM
, and .[27] Variable kernel density estimates and variable kernel density estimates. Aust. J. Stat. 32 (1990) 361–371. | DOI | MR
.[28] Approximately sufficient statistics and Bayesian computation. Stat. Appl. Genet. Mol. Biol. 7 (2008) Art. ID 26. | MR | Zbl
and .[29] On conditional distributions of nearest neighbors. J. Multivariate Anal. 42 (1992) 67–76. | DOI | MR | Zbl
and .[30] A nonparametric estimate of a multivariate density function. Ann. Math. Statist. 36 (1965) 1049–1051. | DOI | MR | Zbl
and .[31] Multivariate -nearest neighbor density estimates. J. Multivariate Anal. 9 (1979) 1–15. | DOI | MR | Zbl
and .[32] Bayesian Core: A Practical Approach to Computational Bayesian Statistics. Springer, New York, 2007. | MR | Zbl
and .[33] Relevant statistics for Bayesian model choice. J. R. Stat. Soc. Ser B. To appear, 2014. | MR | Zbl
, , and .[34] Approximate Bayesian computational methods. Stat. Comput. 22 (2012) 1167–1180. | DOI | MR | Zbl
, , and .[35] Equations of state calculations by fast computing machines. J. Chem. Phys. 21 (1953) 1087–1091. | DOI | Zbl
, , , and .[36] Consistency properties of nearest neighbor density function estimators. Ann. Statist. 5 (1977) 143–154. | DOI | MR | Zbl
and .[37] Large sample properties of nearest neighbor density function estimators. In Statistical Decision Theory and Related Topics II: Proceedings of a Symposium Held at Purdue University, May 17–19, 1976, S. S. Gupta and D. S. Moore (Eds) 269–279. Academic Press, New York, 1977. | MR | Zbl
and .[38] On estimating regression. Theory Probab. Appl. 9 (1964) 141–142. | DOI | Zbl
.[39] On nonparametric estimates of density functions and regression curves. Theory Probab. Appl. 10 (1965) 186–190. | DOI | MR | Zbl
.[40] On the estimation of a probability density function and the mode. Ann. Math. Statist. 33 (1962) 1065–1076. | DOI | MR | Zbl
.[41] Population growth of human Y chromosomes: A study of Y chromosome microsatellites. Mol. Biol. Evol. 16 (1999) 1791–1798. | DOI
, , and .[42] Stochastic Simulation. Wiley, New York, 1982. | MR | Zbl
.[43] Monte Carlo Statistical Methods, 2nd edition. Springer, New York, 2004. | MR | Zbl
and .[44] Lack of confidence in approximate Bayesian computation model choice. Proc. Natl. Acad. Sci. USA 108 (2011) 15112–15117. | DOI
, , and .[45] Conditional probability density and regression estimates. In Multivariate Analysis II, P. R. Krishnaiah (Ed.) 25–31. Academic Press, New York, 1969. | MR
.[46] A class of non-parametric estimates of a smooth regression function. Ph.D. thesis, Stanford Univ., 1966. | MR
.[47] Bayesianly justifiable and relevant frequency calculations for the applied statistician. Ann. Statist. 12 (1984) 1151–1172. | DOI | MR | Zbl
.[48] Sequential Monte Carlo without likelihoods. Proc. Natl. Acad. Sci. USA 104 (2007) 1760–1765. | DOI | MR | Zbl
, and .[49] Singular Integrals and Differentiability Properties of Functions. Princeton Univ. Press, Princeton, 1970. | MR | Zbl
.[50] Consistent nonparametric regression. Ann. Statist. 5 (1977) 595–645. | DOI | MR | Zbl
.[51] Inferring coalescence times from DNA sequence data. Genetics 145 (1997) 505–518. | DOI
, , and .[52] Smooth regression analysis. Sankhya A 26 (1964) 359–372. | MR | Zbl
.[53] Measure and Integral. An Introduction to Real Analysis. Marcel Dekker, New York, 1977. | MR | Zbl
and .[54] Approximate Bayesian computation (ABC) gives exact results under the assumption of model error. Stat. Appl. Genet. Mol. Biol. 12 (2008) 129–141. | MR
.[55] Trigonometric Series, Vol. II. Cambridge Univ. Press, Cambridge, 1959. | MR | Zbl
.Cited by Sources: