A Variational Bayesian Approach for Estimating Parameters of a Mixture of Erlang Distribution

This article proposes a fast algorithm for estimating parameters of a mixture of Erlang (MER) distribution from empirical samples. In particular, we develop a variational approach for approximately computing posterior distributions of parameters of MER distribution in the Bayesian context. Computation speed of the proposed method becomes up to 200 times faster than that of the Markov chain Monte Carlo (MCMC) method. The estimates of proposed method are almost same as those of MCMC method. Moreover, we discuss how to estimate shape parameters of Erlang distribution based on a certain goodness-of-fit criterion in the proposed variational Bayes method.

[1]  Marcel F. Neuts,et al.  Matrix-Geometric Solutions in Stochastic Models , 1981 .

[2]  Jeff A. Bilmes,et al.  A gentle tutorial of the em algorithm and its application to parameter estimation for Gaussian mixture and hidden Markov models , 1998 .

[3]  M. Malhotra,et al.  Selecting and implementing phase approximations for semi-Markov models , 1993 .

[4]  Peter Buchholz,et al.  A Novel Approach for Phase-Type Fitting with the EM Algorithm , 2006, IEEE Transactions on Dependable and Secure Computing.

[5]  M. A. Johnson,et al.  Selecting Parameters of Phase Distributions: Combining Nonlinear Programming, Heuristics, and Erlang Distributions , 1993, INFORMS J. Comput..

[6]  Axel Thümmler,et al.  Efficient phase-type fitting with aggregated traffic traces , 2007, Perform. Evaluation.

[7]  Hagai Attias,et al.  Inferring Parameters and Structure of Latent Variable Models by Variational Bayes , 1999, UAI.

[8]  Naonori Ueda,et al.  Bayesian model search for mixture models based on optimizing variational bounds , 2002, Neural Networks.

[9]  Steve R. Waterhouse,et al.  Bayesian Methods for Mixtures of Experts , 1995, NIPS.

[10]  Nozer D. Singpurwalla,et al.  Reliability and risk , 2006 .

[11]  G. McLachlan,et al.  The EM algorithm and extensions , 1996 .

[12]  Subir Ghosh,et al.  Reliability and Risk: A Bayesian Perspective , 2008, Technometrics.

[13]  C. Robert,et al.  Estimation of Finite Mixture Distributions Through Bayesian Sampling , 1994 .

[14]  Ramin Sadre,et al.  Fitting World Wide Web request traces with the EM-algorithim , 2001, SPIE ITCom.

[15]  R. Miller,et al.  Bayesian Analysis of the Two-Parameter Gamma Distribution , 1980 .

[16]  W. Wong,et al.  The calculation of posterior distributions by data augmentation , 1987 .

[17]  New York Dover,et al.  ON THE CONVERGENCE PROPERTIES OF THE EM ALGORITHM , 1983 .

[18]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[19]  C. Robert,et al.  Computational and Inferential Difficulties with Mixture Posterior Distributions , 2000 .

[20]  A. Bobbio,et al.  A benchmark for ph estimation algorithms: results for acyclic-ph , 1994 .

[21]  Michael A. Johnson,et al.  Matching moments to phase distri-butions: mixtures of Erlang distribution of common order , 1989 .

[22]  Tadashi Dohi,et al.  Hyper-Erlang Software Reliability Model , 2008, 2008 14th IEEE Pacific Rim International Symposium on Dependable Computing.

[23]  Yukito Iba.  EXTENDED ENSEMBLE MONTE CARLO , 2001 .