. "GA201/98/0478" . . "0"^^ . . . . "Struktury podm\u00EDn\u011Bn\u00E9 nez\u00E1vislosti: informa\u010Dn\u011B-teoretick\u00FD p\u0159\u00EDstup" . "http://www.isvav.cz/projectDetail.do?rowId=GA201/98/0478"^^ . "Pojem podm\u00EDn\u011Bn\u00E9 nez\u00E1vislosti je podstatn\u00FD pro mnoho ot\u00E1zek spojen\u00FDch s rozhodov\u00E1n\u00EDm za nejistoty, zejm\u00E9na v oblasti tzv. pravd\u011Bpodobnostn\u00EDho rozhodov\u00E1n\u00ED (ale i v r\u00E1mci jin\u00FDch kalkul\u016F nejistoty). Z\u00E1m\u011Brem projektu je \u0159e\u0161it n\u011Bkter\u00E9 matematick\u00E9 probl\u00E9my, kter\u00E9 vyvst\u00E1vaj\u00ED v souvislosti s form\u00E1ln\u00EDmi vlastnostmi struktur podm\u00EDn\u011Bn\u00E9 nez\u00E1vislosti a v souvislosti se zp\u016Fsoby reprezentace t\u011Bchto struktur v po\u010D\u00EDta\u010Di. Uve\u010Fme 3 konkr\u00E9tn\u00ED c\u00EDle. Za prv\u00E9 zjistit, zda je mo\u017En\u00E9 odvodit dal\u0161\u00ED informa\u010Dn\u011B-teoretick\u00E9 nerovnostipro tzv. entropickou funkci (o t\u011Bchto nerovnostech je zn\u00E1mo, \u017Ee maj\u00ED hlubokou souvislost s form\u00E1ln\u00EDmi vlastnostmi podm\u00EDn\u011Bn\u00E9 nez\u00E1vislosti). Za druh\u00E9 prozkoumat podrobn\u011Bji pojem slo\u017Eitosti semigrafoidu (co\u017E slibuje efekt\u00EDvn\u011Bj\u0161\u00ED zp\u016Fsob reprezentace nez\u00E1vis lostn\u00EDch struktur v po\u010D\u00EDta\u010Di). Za t\u0159et\u00ED, porovnat vlastnosti r\u016Fzn\u00FDch kvantitativn\u00EDch m\u011Br z\u00E1vislosti (co\u017E slibuje mo\u017Enost odhadu nez\u00E1vislostn\u00EDch struktur na z\u00E1klad\u011B dat)." . "7"^^ . . . "7"^^ . . . "Byly z\u00EDsk\u00E1ny teoretick\u00E9 v\u00FDsledky ve statistice a teorii informace. V\u00FDstupy (publikace) jsou odpov\u00EDdaj\u00EDc\u00ED d\u00E9lce trv\u00E1n\u00ED grantu. Lze p\u0159edpokl\u00E1dat, \u017Ee v p\u0159\u00EDpad\u011B dodr\u017Een\u00ED pravidel, by grant \u00FAsp\u011B\u0161n\u011B pokra\u010Doval dal\u0161\u00ED 2 roky (nebyla v\u010Das dod\u00E1na v\u00FDro\u010Dn\u00ED zpr\u00E1va)."@cs . . . "1"^^ . . . "The concept of conditional independence is essential for many tasks connected with decision - making under uncertainty , especially in the area of probabilistic reasoning , but also in the framework of other uncertainty calculi. The aim of the project is to solve several mathematical problems which arise in connection with formal properties of conditional independence structures and in connection with methods of computer representation of these structures. Let us mention 3 specific goals. First , to find whether it is possible to derive further information - theoretical inequalities for entropic functions ( it is known that these inequalities have deep connection with formal properties of conditional independence structures ). Second, to explore the concept of complexity of a semigraphoid ( which promises a more effective way of computer representation of condional independence structures ). Third , to compare properties of various quantitative measures of dependence ( which promises the possibility"@en . "Conditional independence structures: information-theoretical approach"@en . "0"^^ . . .