03561ntm a22005297i 4500 000696928 CZ-PrVSE 20230527092730.0 m d cr n|||||||||| 230527s2023 xr fsbm 000 0 eng d NEZPRACOVANÝ IMPORT ABA006 cze ABA006 ABA006 rda Zeman, Václav ISIS:96325 dis Dolování pravidel nad propojenými daty eng Rule mining over linked data / Václav Zeman 2023 ?? stran : digital, PDF soubor Vedoucí práce: Tomáš Kliegr Disertační práce (Ph.D.)—Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky, 2023 Obsahuje bibliografii Textový (vysokoškolská kvalifikační práce) Rok obhajoby 2023 Rule learning is one way for knowledge discovery from databases. With the growing amount of machine-readable and linked data published on the web, conventional rule mining algorithms needed to be adapted to the properties of the knowledge graphs into which linked data are grouped. Knowledge graphs can contain millions of statements, describe different entities in different contexts, have a high degree of interconnectedness, are considered incomplete, and are therefore open to further dissemination. One of the algorithms capable of efficiently and quickly mining rules from knowledge graphs is AMIE. This algorithm combines various techniques from the field of inductive logic programming and association rule mining. This thesis presents several algorithmic extensions to AMIE, which make it faster and more robust. All proposed approaches have been implemented and encapsulated in a single rule mining system called RDFRules, which can handle many knowledge graphs at once, merge them and analyze input data, mine rules according to the AMIE approach, process output rules, and use them for predictive or descriptive purposes. The performed evaluations show that the proposed improvements of the AMIE algorithm reduce the mining time by up to several orders of magnitude. Furthermore, it has been shown that the newly proposed data pre-processing and rules post-processing methods can discover additional knowledge and provide a better predictive ability of the assembled models than in the case of other common approaches. Způsob přístupu: Internet aplikovaná informatika [obor disert. práce] disertace fd132024 czenas dissertations eczenas rule learning data mining knowledge graphs linked data Kliegr, Tomáš ISIS:8484 ths Popelínský, Lubomír opn Vysoká škola ekonomická v Praze. Fakulta informatiky a statistiky kn20010709399 dgg https://insis.vse.cz/zp/53767/podrobnosti VŠKP v InSIS https://insis.vse.cz/zp/53767 Hlavní práce https://insis.vse.cz/zp/53767/posudek/vedouci Hodnocení vedoucího https://insis.vse.cz/zp/53767/posudek/oponent/77737 Oponentura https://insis.vse.cz/zp/53767/posudek/oponent/77738 Oponentura https://insis.vse.cz/zp/53767/posudek/oponent/77740 Oponentura https://insis.vse.cz/zp/53767/podrobnosti dc:identifier NEPOSILAT VSKP vse53767 230526 53767