Stvaranje naredbi jezika SQL iz prirodnog jezika

master's thesis

• Title (english): Generating SQL Instructions from Natural Language
• Author: Mihael Matijaščić
• Mentor: Goran Delač (mentor)
• Committee member: Goran Delač (predsjednik povjerenstva)
• Committee member: Klemo Vladimir (član povjerenstva)
• Committee member: Siniša Srbljić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

ovom diplomskom radu predstavlja se problem pretvaranja korisniˇckog upita
prema bazi podataka, napisanog engleskim jezikom, u SQL upit (skra´ceno "Tekst u
SQL"). U prvom dijelu rada objašnjavaju se teoretski pojmovi vezani uz navedeni problem.
Stoga, spominje se sama obrada ulaznog teksta, te njegova priprema za ulaz u
arhitekturu modela koji rješava problem "Text to SQL". Zatim se objašnjavaju teoretski
pojmovi vezani uz strukturu modela, kao što su ponavljaju´ce neuronske mreže, te
razne varijante takvih mreža. Na kraju prikazuje se primjer jednog modela koji koristi
WikiSQL skup podataka za rješavanje problema "Text to SQL"„ te se predstavlja programsko
rješenje koje predani korisniˇcki upit iz WikiSQL skupa podataka pretvara u
SQL upit pomo´cu prethodnog modela.

Abstract (english)

This thesis presents the problem of converting a user query according to a database,
written in English, into an SQL query (in short textit "Text in SQL"). The first part
of the paper explains the theoretical concepts related to this problem. Therefore, the
processing of the input text itself is mentioned, as well as its preparation for entering
the architecture of the model that solves the problem textit "Text to SQL". Then, theoretical
concepts related to the structure of the model are explained, such as repetitive
neural networks, and various variants of such networks. Finally, an example of a model
that uses the WikiSQL dataset to troubleshoot textit "Text to SQL" is presented, and a
software solution is presented that converts a submitted user query from the WikiSQL
dataset into an SQL query using the previous model.

• Keywords: SQL; model; tekst; upit; ponavljajuće neuronske mreže; analiza ulaznog teksta; BERT; obrada prirodnog jezika
• Keywords (english): SQL; model; text; question; query; recurrent neural networks; analysis of input text; BERT; natural language processing.
• Language: croatian
• URN:NBN: urn:nbn:hr:168:317410
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 15:34:25