Examining the Trend of Literature on Classification Modelling: A Bibliometric Approach

: This paper analyses and reports various types of published works related to classification or discriminant modelling. This paper adopted a bibliometric analysis based on the data obtained from the Scopus online database on 27th July 2019. Based on the ‘keywords’ search results, it yielded 2775 valid documents for further analysis. For data visualisation purposes, we employed VOSviewer. This paper reports the results using standard bibliometric indicators, particularly on the growth rate of publications, research productivity, analysis of the authors and citations. The outcomes revealed that there is an increased growth rate of classification literature over the years since 1968. A total of 2473 (89.12%) documents were from journals ( n =1439; 51.86%) and conference proceedings ( n =1034; 37.26%) contributed as the top publications in this classification topic. Meanwhile, 2578 (92.9%) documents are multi-authored with an average collaboration index of 3.34 authors per article. However, this classification research field found that the famous numbers of authors’ collaboration in a document are two (with n =758; 27.32%), three ( n =752; 27.10%) and four ( n =560; 20.18%) respectively. An analysis by country, China with 1146 (41.30%) published documents thus is ranked first in productivity. With respect to the frequency of citations, Bauer and Kohavi (1999)’s article emerged as the most cited article through 1414 total citations with an average of 70.7 citations per year. Overall, the increasing number of works on classification topics indicates a growing awareness of its importance and specific requirements in this research field.