An Application of Adaptive Independent Chain Metropolis–Hastings Algorithms in Bayesian Hazard Rate Estimation

Gõsemyr, Jørund; Natvig, Bent; Nygõrd, Cecilie Sjuls
September 2004
Methodology & Computing in Applied Probability;Sep2004, Vol. 6 Issue 3, p293
Academic Journal
In this paper, we consider adaptive independent chain (AIC) Metropolis-Hastings algorithms as introduced in a special context in Gåsemyr et al. (2001) and developed theoretically in Gåsemyr (2003). The algorithms aim at producing samples from a specific target distribution Π, and are adaptive, non-Markovian versions of the Metropolis-Hastings independent chain. A certain parametric class of possible proposal distributions is fixed, and the parameters of the proposal distribution are updated periodically on the basis of the recent history of the chain, thereby obtaining proposals that get ever closer to Π In the former paper a version of these algorithms was shown to be very efficient compared to standard simulation techniques when applied to Bayesian inference in reliability models with at most three dependent parameters. The aim of the present paper is to investigate the performance of the AIC algorithm when the number of dependent parameters and the complexity of the model increases. As a test case we consider a model treated in Arjas and Gasbarra (1996). The target distribution Π is the posterior distribution for the vector X = (X1,…,Xn) of dependent parameters, representing a piecewise constant approximation to the hazard rate X(t), where t0 ≤ t ≤ tn. Especially, for the case n = 12 it turned out that some versions of the AIC were very efficient compared to standard simulation techniques and also to the algorithm applied in Arjas and Gasbarra (1996). This includes a version of the componentwise adaptive independent chain the basic idea of which was given in Gåsemyr (2003).


Related Articles

  • Statistics unravelled. Logsdon, John // Scientific Computing World;Dec2005/Jan2006, Issue 85, p50 

    This article discusses maximum likelihood (ML) and shows how it can be used, with the Weibull distribution, to predict a 1% time to failure. Statisticians use ML to imply the most likely outcome. It is a powerful tool for fitting the model to the data. Ergo, likelihood is a function of the...

  • ENHANCING THE PERFORMANCE OF THE DNDP ALGORITHM. Al-Bahadili, Hussein; Issa, Ghassan; Sabri, Alia // International Journal of Wireless & Mobile Networks;Apr2011, Vol. 3 Issue 2, p113 

    A novel dynamic noise-dependent probabilistic (DNDP) route discovery algorithm was recently developed to enhance the performance of the dynamic probabilistic algorithm in noisy mobile ad hoc networks (MANETs). In this algorithm, the mathematical model for calculating node retransmission...

  • PRIOR ELICITATION FOR A BINOMIAL PROPORTION. Arnold, Barry C. // Pakistan Journal of Statistics;2009, Vol. 25 Issue 4, p419 

    Prior elicitation will always be imprecise, unless the prior is selected from some convenience family of priors. In addition, it is likely that elicited prior information will be incompatible. Possible approaches to resolving the problems associated with imprecise/and or incompatible prior...

  • Synthesis of Linear Array of Parallel Dipole Antennas with Minimum Standing Wave Ratio Using Simulated Annealing and Particle Swarm Optimization approach. Basu, Banani; Mahanti, G. K. // International Journal on Computer Science & Engineering;2010, Vol. 2 Issue 3, p818 

    In this paper, we propose a technique based on two evolutionary algorithms simulated annealing and particle swarm optimization to design a linear array of half wavelength long parallel dipole antennas that will generate a pencil beam in the horizontal plane with minimum standing wave ratio (SWR)...

  • Periodical Replacement Model for a Multi-Unit System Subject to Failure Rate Interaction. Lai, Min-Tsai // Quality & Quantity;Jun2007, Vol. 41 Issue 3, p401 

    A periodical replacement model for a multi-unit system subject to failure rate interaction between units is presented. In a N-unit system, one is dominant unit and the others are secondary units. When a secondary unit fails, it increases the failure rate of the dominant unit by some amount,...

  • A PROOF OF THE SHAPE OF THE BIRNBAUM-SAUNDERS HAZARD RATE FUNCTION. Bebbington, Mark; Chin-Diew Lai; Zitikis, Ricardas // Mathematical Scientist;Jun2008, Vol. 33 Issue 1, p49 

    We prove that the Glaser's η-function of the Birnbaum-Saunders distribution is upside- down bathtub-shaped. This implies - as a consequence of Glaser's general result - that the Birnbaum-Saunders hazard rate function is upside-down bathtub-shaped. To the best of our knowledge, this is the...

  • FAILURE-RATE FUNCTION.  // Encyclopedia of Operations Research & Management Science;2001, p288 

    An encyclopedia entry about "failure-rate function" is presented. The failure rate function is defined by the approximate probability that a random-lifetime ends in a small interval of time. Other terms for function h(t) are presented.

  • Testing Inference in Accelerated Failure Time Models. Medeiros, Francisco M. C.; da Silva-Júnior, Antônio H. M.; Valenća, Dione M.; Ferrari, Silvia L. P. // International Journal of Statistics & Probability;May2014, Vol. 3 Issue 2, p121 

    We address the issue of performing hypothesis testing in accelerated failure time models for non-censored and censored samples. The performances of the likelihood ratio test and a recently proposed test, the gradient test, are compared through simulation. The gradient test features the same...

  • Unobserved heterogencity in hazard rate models: a test and an illustration from a study of career mobility. Blossfeld, Hans-peter; Hamerle, Alfred // Quality & Quantity;Jun89, Vol. 23 Issue 2, p129 

    This paper proposes a test for neglected heterogeneity in continuous-time hazard rate models that can be done easily using generally available program packages. It is a score test appropriate for eases in which the variance of the heterogeneity is small and it can be applied quite generally,...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics