"Pro dob\u00FDv\u00E1n\u00ED znalost\u00ED z rela\u010Dn\u00EDch dat neexistuje zat\u00EDm \u017E\u00E1dn\u00E1 metoda umo\u017E\u0148uj\u00EDc\u00ED sou\u010Dasn\u011B robustn\u00ED reprezentaci a inferenci znalost\u00ED na \u00FArovni logiky 1. \u0159\u00E1du a pr\u00E1ci s nep\u0159esnost\u00ED v datech. Tyto probl\u00E9my jsou dosud \u0159e\u0161eny separ\u00E1tn\u011B. \u00DAkolem projektu je navrhnout spojen\u00ED existuj\u00EDc\u00EDch reprezenta\u010Dn\u00EDch r\u00E1mc\u016F pro oba t\u00E9matick\u00E9 okruhy s c\u00EDlem z\u00EDskat, implementovat a odladit efektivn\u00ED n\u00E1stroj pro pr\u00E1ci s komplexn\u00ED, ale neur\u010Ditou expertn\u00ED znalost\u00ED pro \u00FA\u010Dely n\u00E1sledn\u00E9 automatick\u00E9 konstrukce hypot\u00E9z. \u0158e\u0161en\u00ED projektu bude navazovat a d\u00E1le rozv\u00EDjet p\u016Fvodn\u00ED teoretick\u00E9 v\u00FDsledky navrhuj\u00EDc\u00EDho t\u00FDmu t\u00FDkaj\u00EDc\u00ED se slou\u010Den\u00ED pravd\u011Bpodobnostn\u00ED inference s rela\u010Dn\u00ED logikou. Jako inspirativn\u00ED prost\u0159ed\u00ED pro testov\u00E1n\u00ED metodiky a n\u00E1stroj\u016F poslou\u017E\u00ED p\u016Fvodn\u00ED medic\u00EDnsk\u00E1 data neuniformn\u00EDho charakteru, pro jejich\u017E interpretaci je t\u0159eba vyu\u017Eit\u00ED komplexn\u00ED expertn\u00ED znalosti formulovan\u00E9 v t\u011Bsn\u00E9 spolupr\u00E1ci s pracovi\u0161t\u011Bm spolunavrhovatele. V\u00FDsledky anal\u00FDz p\u0159isp\u011Bj\u00ED k pochopen\u00ED slo\u017Eit\u00FDch biomedic\u00EDnsk\u00FDch d\u011Bj\u016F, nap\u0159. genov\u00E9 exprese." . . . " knowledge reprezentation" . . "2009-12-31+01:00"^^ . "2010-04-15+02:00"^^ . . . . "Relational machine learning for analysis of biomedical data"@en . "Rela\u010Dn\u00ED strojov\u00E9 u\u010Den\u00ED pro pr\u016Fzkum biomedic\u00EDnsk\u00FDch dat" . "http://www.isvav.cz/projectDetail.do?rowId=1ET101210513"^^ . . "53"^^ . "53"^^ . "Byly navr\u017Eeny a testov\u00E1ny n\u00E1stroje pro konstrukci dom\u00E9nov\u00E9 znalosti kombinuj\u00EDc\u00ED pravd\u011Bpodobnostn\u00ED inferenci s rela\u010Dn\u00ED logikou. Byl vytvo\u0159en SW n\u00E1stroj pro tvorbu datab\u00E1ze z l\u00E9ka\u0159sk\u00FDch zpr\u00E1v, kter\u00E9 umo\u017Enily vznik a anal\u00FDzu datab\u00E1ze NF1."@cs . "1ET101210513" . . " multicriterial decisison making" . . . . "relational machine learning" . "2009-03-11+01:00"^^ . "1"^^ . "2"^^ . . . " domain knowledge" . "0"^^ . . . "relational machine learning; multicriterial decisison making; knowledge reprezentation; domain knowledge; data mining"@en . "There is no data mining method enabling both for robust knowledge representation/inference on the level of 1st order logic and for uncertainty in the treated data since both types of problems are treated separately. Our project aims to join existing representation frameworks with intention to design and implement an efficient tool for automated construction of hypothesis about the data utilizing complex but uncertain expert knowledge. Suggested methodology builds on the original theretical results of the proposer's team representing an attempt to merge probabilistic inference with relational logics. The designed approach/ tools will be verified when analysing original medical data inherently lacking uniform structure and requiring complex expert knowledge for their interpretation. This knowledge will be specified in close cooperation with the co-proposing team of medical faculty. The data analysis results will improve understanding of complex biomedical processes, eg. gene expression."@en . "There were designed and tested tools for construction of domain knowledge combining probabilistic inference with relational logic. There was created SW tool for building a database from clinical records that allowed to build and analyze NF1 database."@en . . "2005-01-01+01:00"^^ . . . . .