2022 Sažetak: Obradom prirodnog jezika omogućava se da računari uspostave komunikaciju sa čovekom, da razumeju jezik čoveka, ali i da ga transformišu i da se čoveku obrate na njemu prirodnom jeziku. Problem istinitnosne vrednosti iskaza izrečenih prirodnim jezikom je teško utvrditi zbog bogatstva rečnika i dvosmislenosti značenja. Preslikavanjem iskaza prirodnog jezika u iskaze predikatske logike moguće je utvrditi odnose među različitim predikatskim iskazima kao i izvršiti transformaciju iz jednog oblika pojavljivanja u drugi. Jedna od primena ovakvog znanja može biti u automatizovanom kreiranju testova znanja. Abstract: By processing natural language, we enable computers to establish communication with humans, to understand human language, but also to transform it and to address human in natural language. The problem of the truth value of statements uttered in natural language is difficult to determine due to the richness of vocabulary and ambiguity of meaning. By mapping the statements of natural language into the statements of predicate logic, it is possible to determine the relationships between different predicate statements as well as to perform the transformation from one form of appearance to another. One of the applications of this knowledge can be in the automated creation of knowledge tests. srp https://phaidrabg.bg.ac.rs/o:35323 ISBN: 978‐86‐7466‐930‐3 Sistem za automatizaciju testova za proveru znanja baziran na transformaciji predikatskih iskaza System for automation of knowledge verification tests based on transformation of predicate statements application/pdf 1358316 bytes ZBORNIK RADOVA, LXVI KONFERENCIJA ETRAN, Novi Pazar 6 - 9. juna 2022. startpage: VI1.5-1 endpage: VI1.5-2 info:eu-repo/semantics/article All rights reserved Ključne reči: obrada prirodnog jezika, predikatska logika, automatizovani testovi znanja, srpski jezik Marovac, Ulfeta Avdić, Aldina