A Bibliometric Analysis of Highly Cited Artificial Intelligence Publications in Science Citation Index Expanded

Yuh-Shan Ho,Juan-Jose Prieto-Gutierrez
2024-11-15
Abstract:This study aimed to identify and analyze the characteristics of highly cited publications in the field of artificial intelligence within the Science Citation Index Expanded from 1991 to 2022. The assessment focused on documents that garnered 100 citations or more from the Web of Science Core Collection, spanning from their publication date to the end of 2022. Various aspects of these documents were analyzed, encompassing document types, the distribution of annual production, the average number of citations per publication, Web of Science categories, and journals. Moreover, the publication performance of countries, institutions, and authors underwent evaluation through six publication indicators and associated citation metrics. To facilitate a comprehensive comparison of the authors research performance, the Y-index was employed. The outcomes of the analysis revealed that a majority of the highly cited articles were published within the Web of Science categories of "artificial intelligence computer science" and "electrical and electronic engineering". Notably, the United States exhibited dominance across all six publication indicators. Within the realm of average citations per publication, the United Kingdom emerged as a leader for independent articles, first-author articles, and corresponding-author articles. Exceptionally, the Chinese Academy of Sciences in China and the Massachusetts Institute of Technology (MIT) in the USA, contributed significantly. The significant impact of highly cited articles extended to the output of Stanford University in the USA. B.L. Bassler published the most highly cited articles. Upon employing the Y-index analysis, J.E.P. Santos was identified as having the highest potential for publication. In addition to the primary analysis, this study also presented nine classic articles that have left an indelible mark on artificial intelligence research.
Digital Libraries
What problem does this paper attempt to address?
This paper aims to identify and analyze the characteristics of highly - cited artificial intelligence publications included in the Science Citation Index Expanded (SCI - EXPANDED) from 1991 to 2022. Specifically, the research focuses on those documents that have received 100 or more citations in the Web of Science Core Collection by the end of 2022 since their publication. The research analyzes multiple characteristics of these documents, including document types, annual output distribution, average number of citations per document, Web of Science categories and journals. In addition, the research also evaluates the publishing performance of countries, institutions and individual authors through six publishing indicators and related citation indicators. In order to comprehensively compare the research performance of authors, the Y - index is adopted in the research. The research results reveal that most highly - cited articles are concentrated in the two Web of Science categories of "Artificial Intelligence Computer Science" and "Electrical and Electronic Engineering", and the United States performs well in these six publishing indicators. In terms of the average number of citations per document, the United Kingdom is in the leading position in independent articles, first - author articles and corresponding - author articles. The Chinese Academy of Sciences and the Massachusetts Institute of Technology have made significant contributions to the output of highly - cited articles. In addition, the research also lists nine classic articles that have a far - reaching impact on artificial intelligence research.