Institutional Research Information Service
UCL Logo
Please report any queries concerning the funding data grouped in the sections named "Externally Awarded" or "Internally Disbursed" (shown on the profile page) to your Research Finance Administrator. Your can find your Research Finance Administrator at https://www.ucl.ac.uk/finance/research/rs-contacts.php by entering your department
Please report any queries concerning the student data shown on the profile page to:

Email: portico-services@ucl.ac.uk

Help Desk: http://www.ucl.ac.uk/ras/portico/helpdesk
Publication Detail
Reliability based Bayesian inference for probabilistic classification: An overview of sampling schemes
  • Publication Type:
  • Authors:
    Byrnes PG, DiazDelaO FA
  • Publication date:
  • Pagination:
    250, 263
  • Published proceedings:
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
  • Volume:
    10630 LNAI
  • ISBN-13:
  • Status:
  • Name of conference:
    37th SGAI International Conference on Artificial Intelligence, AI 2017
  • Conference place:
    Cambridge, UK
  • Conference start date:
  • Conference finish date:
  • Print ISSN:
© Springer International Publishing AG 2017. As physical systems have become more complex, the demand for models which incorporate uncertainties in the systems behavior has grown. Bayesian updating is a powerful method which allows models to be learned as new information and data becomes available. In a high dimensional setting, Bayesian updating requires the computation of an integral which is analytically intractable. Markov Chain Monte Carlo (MCMC) techniques include a popular class of methods to solve such an integral. A disadvantage of MCMC however, is its low computational efficiency for problems with many uncertain parameters. Accordingly, a relation between the Bayesian updating problem and the engineering reliability problem has been established. The BUS (Bayesian Updating with Structural Reliability Methods) approach enables reliability methods which are efficient in high dimensions, but retain the advantages of MCMC, to be applied to Bayesian updating problems. Subset Simulation (SuS) is an efficient Monte Carlo technique suitable for such tasks. The BUS algorithm requires a likelihood multiplier to be calculated prior to implementation of SuS. The issue of correctly choosing a suitable multiplier value has gathered much interest. Consequently, a modified BUS framework has been developed which computes the multiplier automatically. The choice of MCMC algorithm within SuS greatly affects sample quality, model efficiency and sample acceptance rate. A low sample acceptance rate results in many repeated samples and thus low efficiency. As such this research investigates the effect of different sampling schemes within the SuS algorithm on the performance of the modified BUS framework for binary probabilistic classification models.
Publication data is maintained in RPS. Visit https://rps.ucl.ac.uk
 More search options
UCL Researchers
Clinical Operational Research Unit
University College London - Gower Street - London - WC1E 6BT Tel:+44 (0)20 7679 2000

© UCL 1999–2011

Search by