Název:
Discounting or Optimizing? Different Approaches to Pseudo-Belief Function Correction
Autoři:
Daniel, Milan ; Jiroušek, Radim ; Kratochvíl, Václav Typ dokumentu: Příspěvky z konference Konference/Akce: Workshop on Uncertainty Processing - WUPES 2025 /13./, Třešť (CZ), 20250604
Rok:
2025
Jazyk:
eng
Abstrakt: We present and compare several approaches for transforming pseudo-belief functions, constructed from Jeffreys confidence intervals on observational data, into proper belief functions. Two main classes of methods are examined: one based on polyhedral geometry using various optimization strategies, and the other employing generalized belief discounting. Finally, the proposed methods are evaluated on real cybersecurity data and compared with standard upper and lower approximations of pseudo-belief.
Klíčová slova:
belief function; computational geometry; learning Zdrojový dokument: Proceedings of the 13th Workshop on Uncertainty Processing (WUPES’25), ISBN 978-80-7378-525-3 Poznámka: Související webová stránka: https://wupes.utia.cas.cz/2025/Proceedings.pdf#page=113