×

Reconciling curvature and importance sampling based procedures for summarizing case influence in Bayesian models. (English) Zbl 1409.62126

Summary: Methods for summarizing case influence in Bayesian models take essentially two forms: (1) use common divergence measures for calculating distances between the full-data posterior and the case-deleted posterior, and (2) measure the impact of infinitesimal perturbations to the likelihood to study local case influence. Methods based on approach (1) lead naturally to considering the behavior of case-deletion importance sampling weights (the weights used to approximate samples from the case-deleted posterior using samples from the full posterior). Methods based on approach (2) lead naturally to considering the local curvature of the Kullback-Leibler divergence of the full posterior from a geometrically perturbed quasi-posterior. By examining the connections between the two approaches, we establish a rationale for employing low-dimensional summaries of case influence obtained entirely via the variance – covariance matrix of the log importance sampling weights. We illustrate the use of the proposed diagnostics using real and simulated data.

MSC:

62H25 Factor analysis and principal components; correspondence analysis
62F15 Bayesian inference
62J20 Diagnostics, and linear inference and regression

Software:

LOWESS; JAGS
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Bradlow, E. T.; Zaslavsky, A. M., Case Influence Analysis in Bayesian Inference, Journal of Computational and Graphical Statistics, 6, 314-331 (1997)
[2] Cancho, V. G.; Ortega, E. M.; Paula, G. A., On Estimation and Influence Diagnostics for Log-Birnbaum-Saunders Student-t Regression Models: Full Bayesian Analysis, Journal of Statistical Planning and Inference, 140, 2486-2496 (2010) · Zbl 1188.62135
[3] Cho, H.; Ibrahim, J. G.; Sinha, D.; Zhu, H., Bayesian Case Influence Diagnostics for Survival Models, Biometrics, 65, 116-124 (2009) · Zbl 1159.62073
[4] Cleveland, W. S., Lowess: A Program for Smoothing Scatterplots by Robust Locally Weighted Regression, The American Statistician, 35, 54-54 (1981)
[5] Cook, R. D., Assessment of Local Influence, Journal of the Royal Statistical Society, 48, 133-169 (1986) · Zbl 0608.62041
[6] Epifani, I.; Maceachern, S. N.; Peruggia, M., Case-Deletion Importance Sampling Estimators: Central Limit Theorems and Related Results, Electronic Journal of Statistics, 2, 774-806 (2008) · Zbl 1320.62046
[7] Escobar, L. A.; Jr, W. Q., Assessing Influence in Regression Analysis with Censored Data, Biometrics, 48, 507-528 (1992) · Zbl 0825.62838
[8] Fahrmeir, L.; Lang, S., Bayesian Inference for Generalized Additive Mixed Models Based on Markov Random Field Priors, Journal of the Royal Statistical Society, 50, 201-220 (2001)
[9] Garay, A. M.; Lachos, V. H.; Bolfarine, H., Bayesian Estimation and Case Influence Diagnostics for the Zero-Inflated Negative Binomial Regression Model, Journal of Applied Statistics, 42, 1148-1165 (2015) · Zbl 1514.62579
[10] Geweke, J., Bayesian Inference in Econometric Models Using Monte Carlo Integration, Econometrica: Journal of the Econometric Society, 57, 1317-1339 (1989) · Zbl 0683.62068
[11] Hawkins, D. M.; Bradu, D.; Kass, G. V., Location of Several Outliers in Multiple-Regression Data Using Elemental Sets, Technometrics, 26, 197-208 (1984)
[12] Kalbfleisch, J. D., Non-Parametric Bayesian Analysis of Survival Time Data, Journal of the Royal Statistical Society, Series B, 40, 214-221 (1978) · Zbl 0387.62030
[13] Lachos, V. H.; Castro, L. M.; Dey, D. K., Bayesian Inference in Nonlinear Mixed-Effects Models Using Normal Independent Distributions, Computational Statistics & Data Analysis, 64, 237-252 (2013) · Zbl 1468.62109
[14] Lavine, M., Local Predictive Influence in Bayesian Linear-Models with Conjugate Priors, Communications in Statistics: Simulation and Computation, 21, 269-283 (1992) · Zbl 0850.62286
[15] Maceachern, S. N.; Peruggia, M., Bayesian Tools for EDA and Model Building: A Brainy Study, Case Studies in Bayesian Statistics, Volume 162 of Lecture Notes in Statistics, 345-362 (2002), New York: Springer, New York
[16] Mcculloch, R. E., Local Model Influence, Journal of the American Statistical Association, 84, 473-478 (1989)
[17] Millar, R. B.; Stewart, W. S., Assessment of Locally Influential Observations in Bayesian Models, Bayesian Analysis, 2, 365-383 (2007) · Zbl 1331.62156
[18] Miller, R.; Halpern, J., Regression with Censored Data, Biometrika, 69, 521-531 (1982) · Zbl 0503.62091
[19] Murray, T. A.; Hobbs, B. P.; Sargent, D. J.; Carlin, B. P., Flexible Bayesian Survival Modeling with Semiparametric Time-Dependent and Shape-Restricted Covariate Effects, Bayesian Analysis, 11, 381-402 (2016) · Zbl 1357.62281
[20] NOAA, Climate at a Glance (2015)
[21] Peña, D.; Tiao, G. C.; Tsay, R. S., A Course in Time Series Analysis (2011), New York: Wiley, New York
[22] Peng, F. C.; Dey, D. K., Bayesian-Analysis of Outlier Problems Using Divergence Measures, Canadian Journal of Statistics-Revue Canadienne De Statistique, 23, 199-213 (1995) · Zbl 0833.62028
[23] Peruggia, M., On the Variability of Case-Deletion Importance Sampling Weights in the Bayesian Linear Model, Journal of the American Statistical Association, 92, 199-207 (1997) · Zbl 0889.62020
[24] Plummer, M., JAGS: A Program for Analysis of Bayesian Graphical Models Using Gibbs Sampling, Proceedings of the 3rd International Workshop on Distributed Statistical Computing (2003)
[25] Van Der Linde, A., Local Influence on Posterior Distributions Under Multiplicative Modes of Perturbation, Bayesian Analysis, 2, 319-332 (2007) · Zbl 1331.62171
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.