This paper presents the application of a recently developed non-parametric predictive inferential approach for multinomial data to the problem of prediction of occurrence of new failure modes. These inferences are in terms of lower and upper probabilities for the next observation. The lower probability of occurrence of a new failure mode is zero in all cases, as the data never suggest strongly that there have to be further failure modes. The main interest is in the upper probability that the next observed failure is caused by a new failure mode.
CoolenF. P.A.AugustinT.Learning from multinomial data: a nonparametric predictive alternative to the imprecise Dirichlet model. In Proceedings of the Fourth International Symposium onImprecise probabilities and their applications (ISIPTA'05) (Eds CozmanF. G.NauR.SeidenfeldT.), pp. 125–134.
2.
CoolenF. P.A.AugustinT.A nonparametric predictive alternative to the imprecise Dirichlet model: the case of a known number of categoriesInt. J. Approximate Reasoning, 2007, in press.
3.
WalleyP.Statistical reasoning with imprecise probabilities, 1991 (Chapman & Hall, London).
4.
WeichselbergerK.The theory of interval-probability as a unifying concept for uncertaintyInt. J. Approximate Reasoning, 2000, 24, 149–170.
5.
WeichselbergerK.Elementare Grundbegriffe einer Allgemeineren Wahr-scheinlichkeitsrechnung I. Intervallwahrscheinlichkeit als Umfassendes Konzept, 2001 (Physika, Heidelberg).
6.
UtkinL. V.CoolenF. P. A.Imprecise reliability: an introductory overview In Computational intelligence in reliability engineering, Vol. 2, New metaheuristics, neural and fuzzy techniques in reliability (Ed. LevitinG.), 2007, ch. 10, pp. 261–306 (Springer, New York).
7.
CoolenF. P. A.On the use of imprecise probabilities in reliabilityQuality Reliability Engng Int., 2004, 20, 193–202.
8.
CoolenF. P. A.On probabilistic safety assessment in case of zero failuresProc. IMechE, Part O: J. Risk and Reliability, 2006, 220, 105–114.
9.
WalleyP.Inferences from multinomial data: learning about a bag of marbles (with discussion)J. R. Statist. Soc. B, 1996, 58, 3–57.
10.
AugustinT.CoolenF. P. A.Nonparametric predictive inference and interval probabilityJ. Statist. Planning Inference, 2004, 124, 251–272.
11.
CoolenF. P.A.NewbyM. J.Bayesian reliability analysis with imprecise prior probabilitiesReliability Engng System Safety, 1994, 43, 75–85.
12.
HillB. M.Posterior distribution of percentiles: Bayes’ theorem for sampling from a populationJ. Am. Statist. Assoc., 1968, 63, 677–691.
13.
De FinettiB.Theory of probability, 2 volumes, 1974 (John Wiley, London).
14.
CoolenF. P. A.Coolen-SchrijnerP.VanK. J.Nonparametric predictive inference in reliabilityReliability Engng System Safety, 2002, 78, 185–193.
15.
CoolenF. P. A.On nonparametric predictive inference and objective BayesianismJ. Logic, Language Inf., 2006, 15, 21–47.
16.
CoolenF. P. A.Low structure imprecise predictive inference for Bayes’ problemStatist. Probability Lett., 1998, 36, 349–357.
17.
BayesT.An essay towards solving a problem in the doctrine of chancesPhil. Trans. R. Soc., 1763, 53, 370–418, 54, 296–325.
18.
GeisserS.Predictive inference: an introduction, 1993 (Chapman & Hall, London).
19.
BernardoJ. M.SmithA. F. M.Bayesian theory, 1994 (John Wiley, Chichester, West Sussex).
20.
FergusonT. S.A Bayesian analysis of some non-parametric problemsAnn. Statist., 1973, 1, 209–230.
