Probabilistic information retrieval from census data based on distribution mixtures

0410441 - UTIA-B 20000157 RIV CZ eng J - Článek v odborném periodiku
Grim, Jiří - Pudil, Pavel - Somol, Petr
Probabilistic information retrieval from census data based on distribution mixtures.
Acta Oeconomica Pragensia. Roč. 8, č. 2 (2000), s. 41-47. ISSN 0572-3043
Grant ostatní: MŠMT(CZ) VS96063; GA AV(CZ) IAA2075703; GA AV(CZ) KSK1075601
Program: IA; KS
Výzkumný záměr: AV0Z1075907
Kód oboru RIV: BB - Aplikovaná statistika, operační výzkum

The possibility of interactive presentation of statistical properties of large and/or confidential databases by means of probabilistic models is based on the probabilistic expert system PES. The probabilistic model can be estimated by means of EM algorithm in the form of a finite distribution mixture to be used as a knowledge base of the probabilistic expert system PES. The final software product can reproduce the properties of the original database without any further access to the original data.
Trvalý link: http://hdl.handle.net/11104/0130530