급부상 연구 주제 탐지 모델 개발 연구

Abstract

1. 선행연구 및 관련 동향 분석 보고서 발간 2. 복합적 ERT 탐지 모델링을 위한 복합적 데이터 기초 조사 및 수집 1) 확보 가능한 논문 단위 월별 이용통계 조사 2) 논문 단위 월별 이용통계 수집기 개발 3. 월 단위 이용량 시계열 데이터 구축 1) 국내 학술논문의 국내 이용량 시계열 데이터 구축: 98,763건 2) 해외 학술논문의 글로벌 이용량 시계열 데이터 구축: 253,431건 4. 시계열분석 기반 변곡점 탐지 모델 개발 1) 국내 유기화학 분야 학술논문 이용량 데이터 대상 시계열분석 요건 검증 2) 세부 주제분야코드(ASJC) 대비 키워드 동시출현 기반 클러스터링 3) CUSUM(Cumulative sum) 검증 및 BIC(Bayesian information criterion)에 기반 시계열구조 변화 탐지

IV. Results of Research and Development 1. Publication of Trend Report 1) Indicators of Research Impact and Scientific Excellence 2) Indicators of Emergence or Growth of Research Area 3) Detection Models of Emerging Research Topics 2. Survey and Collection of Data 1) Availability of Article-Level Usage Statistics in a Monthly Basis ○ Domestic Scholarly Information Platform: ACOMS, NDSL, NTIS, RISS4U - Survey of Construction Status of Article-Level Usage Statistics ○ Overseas Content Providers: Elsevier, ACS, Highwire Press, PLoS - Survey of Construction Status of Article-Level Usage Statistics 2) Development of Crawler of Article-Level Usage Statistics on the Web - Publisher and Journal List whose Article-Level Usage Statistics are Open to Public - Pattern Analysis of Article-Level Metrics Page - Creation of Virtual URL for Harvesting Article-Level Usage Statistics on the Web - Web Crawling Article-Level Usage Statistics 3. Construction of Time Series in a Monthly Basis Use 1) Time Series for Domestic Use of Domestic Articles: 98,763 articles - Period : 2008. 02 ~ 2015 .04 - Document Type: Article, proceeding article, review, editorial et al. - Times cited indexed by KSCD, SCOPUS - Subject classification assigned from SCOPUS(ASJC) 2) Time Series for Global Use of Global Articles: 253,431 articles - 253,431 articles among 1,189,201 articles in 84 titles - Construction of COUNTER-compliant Usage Statistics Database(COUNTER Release 4 Article Report) - Period : 1999. 01 ~ current 4. Modelling for Detection of Structural Changes in Time Series 1) Time Series Inspection for Time Series Analysis 2) Selection of Models for Detection of Structural Changes 3) Article Clustering for Time Series Analysis 4) Detection Results of Structural Changes in Article Use