Analyzing Scientometric Indicators of Journals and Chief Editors: A Case Study in Artificial Intelligence (AI) Domain

This study investigates the relationship between the rankings of artificial intelligence (AI) journals and the chief editor’s scholarly reputations by mining various scientometric data. The associations between these two types of entities are studied with respect to the top AI journals (selected based on Google Scholar ranking) and journals from various quartiles (based on Scimago quartile ranking). Three quantitative reputation metrics (i.e., citation count, h-index, and i10-index) of editor-in-chief (EiC) and four journal ranking metrics (i.e., h5-index, h5-median, the impact factor (IF), and Scimago Journal Rank (SJR)) of journals are considered to find any relationships. To determine the correlation between various pairs of scholarly metrics of EiC and top AI journals, we employ the Spearman and Kendall correlation coefficients. Furthermore, we investigate whether machine learning (ML) classifiers can predict the SJR and IF of journals utilizing EIC’s scholarly reputation metrics. It is observed that the comparative rankings (based on various metrics) of top AI journals do not correlate with the EiC’s scholarly achievements. The high prediction errors of ML classifiers indicate that the EiC’s scholarly indices are not comprehensive enough to build a good model for predicting the IF or SJR of top AI journals. Nevertheless, when AI journals of various qualities are analyzed, we observe that Q1 journals usually have EiCs with a much higher number of citations and h-index compared to the EiCs from the journals from the bottom two quarterlies (Q3 and Q4). The Mann-Whitney U test indicates the differences between the scholarly metrics of EiCs of Q1 journals and journals from Q3 and Q4 are significant. The results imply that while selecting the EiC of a journal, scientometric indices should be considered prudently.