A General Approach to Probabilistic Data Mining

0531047 - ÚTIA 2022 RIV SG eng C - Conference Paper (international conference)
Jiroušek, Radim - Kratochvíl, Václav
A General Approach to Probabilistic Data Mining.
Sensor Networks and Signal Processing. vol. 176. Singapore: Springer, 2021 - (Peng, S.; Favorskaya, M.; Chao, H.), s. 325-340. 2190-3018. ISBN 978-981-15-4916-8.
[Sensor Networks and Signal Processing (SNSP 2019) /2./. Hualien (TW), 19.11.2019-22.11.2019]
Grant - others:GA ČR(CZ) GA19-06569S; Akademie věd - GA AV ČR(CZ) MOST-04-18
Program: GA
Institutional support: RVO:67985556
Keywords : approximation * probability models * conditional independence * decomposition * information content * ambiguity
OECD category: Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
http://library.utia.cas.cz/separaty/2020/MTR/kratochvil-0531047.pdf

The paper describes principles enabling us to express the knowledge hidden in a multidimensional probability distribution - a distribution that is assumed to have generated the input data - into the form legible by humans, into the form expressible in a plain language. The generality of this approach arises from the fact that we do not assume any type of probability distribution. The basic idea is that the analysis of such a multidimensional distribution is, because of its computational complexity, intractable, and therefore we construct its approximation in a form of a decomposable model, which provides an easy interpretation. The process should be controlled by an expert in the field of application, and the presented principles give him instruction, how, using the tools from probability and information theories, to get satisfactory results.
Permanent Link: http://hdl.handle.net/11104/0310093