ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술지 A Method of Trend Analysis using Latent Dirichlet Allocation
Cited 0 time in scopus Download 9 time Share share facebook twitter linkedin kakaostory
저자
황명하, 하수욱, 인민교, 이강찬
발행일
201805
출처
International Journal of Control and Automation, v.11 no.5, pp.173-182
ISSN
2005-4297
출판사
SERSC
DOI
https://dx.doi.org/10.14257/ijca.2018.11.5.15
협약과제
18HE1700, 빅데이터 상호운용성 지원 표준 개발, 이강찬
초록
© 2018 SERSC Australia. Due to the introduction of text mining, studies have been conducted to analyze meaningful topics and trends in document collections. Trend analysis using Latent Dirichlet Allocation (LDA) in text mining is adopted as one of the trend analysis methods with high accuracy. In this paper, we propose a trend analysis method using LDA. The method is composed of 5 steps and the trend analysis is performed by topic using the extracted result combining LDA and Top 10 keywords. By applying our method and LDA to international standards documents, we performed topic modeling and checked the trend of international standards using extracted topics.
KSP 제안 키워드
Analysis method, High accuracy, International standard, Latent dirichlet allocation (lda), Topic Modeling, Trend analysis, document collections, text mining