Supra-Bayesian Approach to Merging of Incomplete and Incompatible Data

0352234 - ÚTIA 2011 RIV CZ eng C - Conference Paper (international conference)
Sečkárová, Vladimíra
Supra-Bayesian Approach to Merging of Incomplete and Incompatible Data.
Decision Making with Multiple Imperfect Decision Makers. Prague: Institute of Information Theory and Automation Academy of Sciences of the Czech Republic, 2010 - (Guy, T.; Karny, M.; Wolpert, D.), s. 1-6. ISBN 978-80-903834-5-6.
[24th Annual Conference on Neural Information Processing Systems. Whistler, B.C. Canada (CA), 06.12.2010-11.12.2010]
R&D Projects: GA ČR GA102/08/0567; GA MŠMT 1M0572
Institutional research plan: CEZ:AV0Z10750506
Keywords : Bayesian decision making * Supra-Bayesian approach * sharing of probabilistic information
Subject RIV: BC - Control Systems Theory
http://library.utia.cas.cz/separaty/2010/AS/seckarova-supra-bayesian approach to merging of incomplete and incompatible data.pdf

In practice we often need to take every available information into account. Unfortunately the pieces of information given by different sources are often incomplete (with respect to what we are interested in) and have different forms. In this work we try to solve the problem of treating such data in order to get an optimal merger of them. We present a systematic and unified way how to combine the pieces of information by using a Supra-Bayesian approach and other mathematical tools, e.g. Kerridge inaccuracy, maximum entropy principle. To show how the proposed method works a simple example is given at the end of the work.
Permanent Link: http://hdl.handle.net/11104/0191792