2021 VOLUME 1 ISSUE 4
RESEARCH ARTICLE
Application of article-level metrics and comparison with journal-level metrics in differentiated document recommendation: An empirical study in artificial intelligence field
AUTHOR
Feifei Wang, Wanzhao Lu, Yusi Hou
ABSTRACT
The measurement indexes of literature include citation frequency, H index, etc. The evaluation of core journals mainly relies on the indexes such as impact factor or comprehensive evaluation method. With the in-depth development of research, these indicators are not comprehensive and accurate for literature and journals. Therefore, according to various literature needs for different researchers, "core literature" in this study was divided into three types: classical, popular and frontier; and measurement system of document value was constructed with comprehensive use of entropy weight method and principal component analysis from the perspective of Article-Level Metrics. In the case study of artificial intelligence (AI), three types of document sets were acquired with the threshold value of specific indicators, and then measured by a combination of multi-index, achieving identification and recommendation of core documents for different research needs. At the same time, this paper further calculates the total score of the journals according to the literature score, and finds that the journal distribution of different types of core literature is quite different. The difference between the ranking of journal scores and the ranking of impact factor after literature classification is relatively large, but the ranking of journal normalized Eigenfactor and the ranking of impact factor are similar. These research directions, loading journals, selected indicators, and temporal effects in three types of core documents were revealed in the study, which can provide a certain reference for promoting scientific research in AI and launching scientific research management services.
KEYWORDS
Article-Level Metrics; Journal-Level Metrics; Core documents; Entropy weight method; Principal component analysis; Artificial intelligence