• DocumentCode
    3064651
  • Title

    A Reliability Verification Test Model Based on Hybrid Bayesian Prior Distribution

  • Author

    Gao, Feng ; Zheng, Xiaoyun ; Liu, Chang

  • Author_Institution
    Dept. of Autom., Univ. of Harbin Eng., Harbin, China
  • fYear
    2012
  • fDate
    23-26 June 2012
  • Firstpage
    770
  • Lastpage
    774
  • Abstract
    Using advantages of priori Bayesian method, a reliability verification test method based on Hybrid Bayesian Prior Distribution was brought forward. The prior distribution of unknown parameters can be obtained by using conjugate prior distribution method. Prior moment method and Maximum entropy method were used respectively to calculate two different groups of parameters, and then two different prior distributions can be obtained. Then confidence factors of the two prior distributions were determined by using the second category maximum likelihood method, and the final distribution can be got by integrating there two group of parameters according to their weight. Instance proved that the prior distribution obtained by this method is more accurate, and can fit better with the real distribution.
  • Keywords
    Bayes methods; maximum entropy methods; maximum likelihood estimation; program testing; program verification; software reliability; statistical distributions; confidence factors; conjugate prior distribution method; hybrid Bayesian prior distribution; maximum entropy method; maximum likelihood method; prior moment method; priori Bayesian method; reliability verification test method; reliability verification test model; unknown parameters; Bayesian methods; Entropy; Moment methods; Software; Software reliability; Testing; Maximum entropy method; Parameter integration; Prior moment method; Priori Bayesian model; reliability verification testing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Sciences and Optimization (CSO), 2012 Fifth International Joint Conference on
  • Conference_Location
    Harbin
  • Print_ISBN
    978-1-4673-1365-0
  • Type

    conf

  • DOI
    10.1109/CSO.2012.173
  • Filename
    6274837