Title Sustav za prevođenje poslovnih opisa u model podataka entiteta i veza Title (english) System for transformation of business specifications into entity-relationship model Author Sabrina Šuman Mentor Alen Jakupović (mentor)Mentor Sanja Čandrlić (komentor)Committee member Mile Pavlić (predsjednik povjerenstva)Committee member Sanda Martinčić-Ipšić (član povjerenstva)Committee member Zdravko Dovedan Han (član povjerenstva)Granter University of Rijeka Defense date and country 2019-11-25, Croatia Scientific / art field, SOCIAL SCIENCES Universal decimal classificationUDC ) 004 - Computer science and technology. Computing. Data processing Abstract Modeliranje podataka složen je proces koji zahtijeva znanje i iskustvo od dizajnera, a čija kvaliteta znatno utječe na izvedbu u narednim fazama razvoja informacijskog sustava. Karakteristike procesa modeliranja (složen, nealgoritamski, zahtijeva znanje i iskustvo dizajnera) čine ga prikladnim područjem za kreiranje sustava temeljenog na znanju (engl. Knowledge Based System – KBS). Ovim se radom prezentira KB sustav za potporu modeliranju podataka koji je baziran na verbalizaciji. Sustav će upotrebljavati znanje eksperata koje je formalizirano primjenom metoda formalnog jezika. Osnovna funkcionalnost sustava jest prevođenje poslovnog ili nekoga drugog opisa danog u kontroliranom prirodnom engleskom jeziku (engl. Controlled natural English – CEN) u model podataka izraženog u formalnom jeziku modela podataka. Za izvođenje svojih osnovnih funkcionalnosti upotrebljavat će kreiranu bazu znanja temeljenu na pravilima prevođenja primjenom metode raspoznavanja uzoraka. Cjelokupna baza znanja sastoji se od niza kreiranih artefakata (opisani teorijski i implementirani programskim kodom) koji su omogućili izgradnju osnovnog dijela sustava. Definirani artefakti jesu gramatike kontroliranog i formalnog jezika modela podataka (EARC jezik, engl. Entities Attributes Relations Cardinalities), kreiranje posebnog leksikona s nizom dodatnih EARC svojstava, sintaksna analiza kroz prepoznavač kojim se ispituje sintaksna ispravnost, niz pravila prevođenja te sustav postavljanja pitanja i dobivanja odgovora (engl. Question-Answering system – QA system) kojim će biti omogućena komunikacija s korisnikom. Na temelju pregleda relevantne literature te analize procesa modeliranja od strane eksperata definirani su osnovni elementi i funkcionalni zahtjevi KB sustava. Kroz prikaz arhitekture sustava detaljno su opisani svi njegovi dijelovi, korisnici te okolina sustava. Opisani su i priloženi osnovni algoritmi sintaksne analize i prevođenja. Prikazani su primjeri prevođenja s detaljnom analizom svakog koraka. Provedeno je testiranje prototipa sustava s analizom rezultata te su potvrđene postavljene hipoteze. Navedena su i ograničenja sustava i pravci budućeg istraživanja te je o njima diskutirano.
Abstract (english) Data modelling is a complex process which requires knowledge and experience from designers, whose quality has a significant impact on performance in the upcoming stages of information system development. The characteristics of the modelling process (complex, non-algorithmic, requiring knowledge and experience of designers) make it a suitable knowledge-based knowledge base (Knowledge-Based System – KBS). This dissertation presents a system based on data modelling support on verbalization. The system will use expert knowledge which is formalized using formal language methods, in verbalized form. The basic functionality of the system is to translate a business or other description given in Controlled Natural English (CEN) into a model of data expressed in the formal language of the data model. To perform its basic functionality, it will use a created knowledge-based approach using the form-recognition method. The entire knowledge base consists of a series of created artefacts (described in theoretical and implemented code) which have enabled the construction of the basic part of the system. The defined artefacts are the grammar of the controlled and formal language data model (EARC language, Entities Attributes Relations Cardinalities), the creation of a special lexicon with a series of additional EARC properties, syntactic analysis through a recognizer which examines syntax correctness, a set of translation rules, and the Question-Answering system(QA) which will enable communication with the user. Based on a review of the relevant literature and the analysis of the modelling process by the experts, the basic elements and functional requirements of the KB system were defined. Throughout the architecture of the system, all its parts, users and system environment are described in detail. Basic syntax analysis and translation algorithms are described and enclosed. Translation examples are shown with a detailed analysis of each step. Testing of a system prototype was carried out and the hypotheses were confirmed. Limitations of the system and directions for future research as well are enclosed and discussed.
Keywords
poslovni opis
strojno prevođenje
verbalizacija
sustav temeljen na znanju
formalni jezici
raspoznavanja uzoraka
konceptualno modeliranje
metoda entiteta i veza
Keywords (english)
business description
machine translation
verbalisation
knowledge-based systems
formal languages
pattern recognition
conceptual modelling
entity-relationship method
Language croatian URN:NBN urn:nbn:hr:195:652886 Promotion 2021 Study programme Title: Informatics Type of resource Text File origin Born digital Access conditions Access restricted to students and staff of home institution Terms of use Repository Repository of the University of Rijeka, Faculty of Informatics and Digital Technologies Created on 2021-04-26 08:22:35
