원문정보
초록
영어
The purpose of this study is to explore the similarities and differences between author keywords and corpus keywords. Author keywords intuitively given by article authors have been used as a method of information retrieval to identify their articles. Corpus keywords are called keywords automatically extracted from abstract subcopora by using statistical measures such as Log-likelihood. Our research started with a research question regarding whether or not corpus keywords can be used in another way for this information retrieval. This study uses 800 offshore industry journal articles and classifies them four sub-parts, each of which contains a five-year period from 1995 to 2014. This paper compares author keywords with corpus keywords to examine whether or not corpus keywords can be another method for information retrieval both by showing the distribution of the types of n-gram compounds shared in author keywords and corpus keywords. It also compares the two different keywords by using a percent ratio of the types of n-gram compounds shared in the article title corpus as a reference point. Finally it explores the statistical relationship between author keywords and corpus keywords by using a chi square test with a percent ratio of the types of all the n-gram compounds used in the article title corpus as a reference point.
목차
1. 서론
2. 선행연구
3. 자료 및 연구방법
3.1. 해양플랜트 학술논문 코퍼스
3.2 연구방법
4. 연구결과
4.1. 저자키워드와 코퍼스키워드의 타입과 n-gram 합성어 비교
4.2. 공유합성어의 연도기간별 비교
4.3. 논문제목코퍼스 기반 저자키워드와 코퍼스키워드의 비교분석
4.4. 저자키워드와 코퍼스키워드 간의 통계적 관련성
5. 결론
참고문헌