A Bayesian significance test of change for correlated observations

• Autorzy: Abdeldjalil Slama
• Języki publikacji: EN

Abstrakty

EN

This paper presents a Bayesian significance test for a change in mean when observations are not independent. Using a noninformative prior, a unconditional test based on the highest posterior density credible set is determined. From a Gibbs sampler simulation study the effect of correlation on the performance of the Bayesian significance test derived under the assumption of no correlation is examined. This paper is a generalization of earlier studies by KIM (1991) to not independent observations.

Słowa kluczowe

EN

autoregressive model   change point   HPD region sets   p-value   Gibbs sampler  

Kategorie tematyczne

• 62F03: Hypothesis testing
• 62F15: Bayesian inference
• 62M10: Time series, auto-correlation, regression, etc.

• Wydawca: Faculty of Mathematics, Computer Science and Econometrics, University of Zielona Góra
• Czasopismo: Discussiones Mathematicae Probability and Statistics
• Rocznik: 2014
• Tom: 34
• Numer: 1-2
• Strony: 51-62

Daty

wydano

2014

otrzymano

2014-05-12

poprawiono

2014-07-23

Twórcy

autor

Abdeldjalil Slama

• Departemnet of Mathematics and Computer Science, University of Adrar, National Road No. 06, Adrar, Algeria
• Departement of Probability and Statistics, USTHB, P.O. Box 32 EL Alia 16111 Bab Ezzouar, Algiers, Algeria

Bibliografia

• [1] M.M. Barbieri and C. Conigliani, Bayesian analysis of autoregressive time series with change points, J. Italian Stat. Soc. 7 (1998) 243-255. doi: 10.1007/BF03178933
• [2] R.L. Brown, J. Durbin and J.M. Evans, Techniques for testing the constancy of regression relationships over time (with discussion), J.R. Statist. Soc. A 138 (1975) 149-63.
• [3] H. Chernoff and S. Zacks, Estimating the current mean distribution which is subjected to change in time, Ann. Math. Statist. 35 (1964) 999-1018. doi: 10.1214/aoms/1177700517
• [4] D. Ghorbanzadeh and R. Lounes, Bayesian analysis for detecting a change in exponential family, Appl. Math. Comp. 124 (2001) 1-15. doi: 10.1016/S0096-3003(00)00029-1
• [5] H.-J. Kim, Change-point detection for correlated observations, Statistica Sinica 6 (1996) 275-287.
• [6] A. Kander and S. Zacks, Test procedure for possible change in parameters of statistical distributions occuring at unknown time point, Ann. Math. Statist. 37 (1966) 1196-1210. doi: 10.1214/aoms/1177699265
• [7] D. Kim, A Bayesian significance test of the stationarity of regression parametres, Biometrika 78 (1991) 667-675. doi: 10.2307/2337036
• [8] D.V. Hinkley, Inference about the change-point in a sequence of random variables, Biometrika 57 (1970) 1-17. doi: 10.2307/2334932
• [9] E.S. Page, Continuous inspection schemes, Biometrika 41 (1954) 100-115.
• [10] E.S. Page, A test for change in a parameter occurring at an unknown point, Biometrika 42 (1955) 523-527. doi: 10.2307/2333009
• [11] A. Sen and M.S. Srivastava, Some one-sided tests for change in level, Technometrics 17 (1975) 61-64. doi: 10.2307/1268001
• [12] D. Siegmund, Boundary Crossing probabilities and statistical applications, Ann. Statist. 14 (1986) 361-404. doi: 10.1214/aos/1176349928
• [13] D. Siegmund, Confidence sets in change point problem, Int. Statist. Rev. 56 (1988) 31-48. doi: 10.2307/1403360
• [14] K.J. Worsley, The power of likelihood ratio and cumulative sum tests for a change in a binomial probability, Biometrika 70 (1983) 455-464. doi: 10.2307/2335560
• [15] K.J. Worsley, Confidence regions and tests for a change-point in a sequence of exponential family random variables, Biometrika 73 (1986) 91-104. doi: 10.2307/2336275

Identyfikatory

DOI

10.7151/dmps1169