구문 패턴과 개체명사전을 이용한 자연어 질의문장의 SQL 자동 변환 방법

Metadata Downloads
Issued Date
Information society can be realized based on efficient collection, management and retrieval of large - scale information. Since the database (DB) is an important and important field in the information society, a system that facilitates DB access to general users can be regarded as an important core technology in the future. However, there is a restriction that general users should have knowledge of DB and business knowledge such as SQL in order to use DB directly. In order to solve the above limitations, researches on NLIDB (Natural Language Interfaces to Database System) are under way. To construct NLIDB, natural language processing technology and processing technology for DBMS (Database Management System) query language should precede. In this paper, I propose a method to convert to SQL statement so that general users without DB and SQL knowledge can search DB through natural language processing of query for the purpose of searching and using DB with general computing performance. The proposed method generates SQL by matching the columns using N-gram and named entity dictionary through natural language query. We analyze the experiment through the proposed method and describe the experimental result.
Alternative Title
Automatic SQL conversion of natural language query sentences using syntax pattern and named entity dictionary
Alternative Author(s)
HeeChul Shin
산업기술융합대학원 소프트웨어융합공학과
Awarded Date
2018. 8
Table Of Contents

I. 서론 1
A. 연구 배경 및 목적 1
B. 연구 내용 및 구성 2

II. 관련연구 3
A. 자연어 처리 3
1. 자연어 처리 개념 3
2. N-gram 4
B. NLIDB(Natural Language Interface for Database) 5
1. NLIDB 개념 5
2. 기존연구 5

III. 자연어 질의문장 SQL 변환 9
A. 전체 시스템 개요 9
B. 자연어 질의문장 전처리 과정 12
C. 의문사구와 명사구 추출 15
1. 의문사구 추출 16
2. 명사구 추출 18
D. SQL 변환 19
1. 의문사구 SQL 변환 19
2. 명사구 SQL 변환 20
E. 개체명 인식(Named Entity Recognition) 21
1. 개체명 사전 21
F. SQL 생성 22
IV. 실험 및 평가 23
A. 평가 28

V. 결론 및 제언 30

참고문헌 31
신희철. (2018). 구문 패턴과 개체명사전을 이용한 자연어 질의문장의 SQL 자동 변환 방법
Appears in Collections:
Engineering > Theses(Master)(산업기술창업대학원)
Authorize & License
  • AuthorizeOpen
Files in This Item:

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.