Abstract:
Bibliometrics is used to determine patterns in published research. The aim of this paper is to illustrate the observable bibliometric patterns in the journal Southern Forests: a Journal of Forest Science. Frequency analysis and co-occurrence network analysis were performed to identify patterns. Natural Language Processing and Supervised Machine Learning were used to perform text classification. The objective of the text classification was to classify articles into 15 themes. Each article was categorised in terms of the two main themes associated with the article. The analysis included 1 574 publications from 1941 to 2020 and confirmed that the journal’s change in name and aims were successful in increasing the number of international researchers publishing in the journal. The research institute co-occurrence network diagram illustrates that there are two main research collaboration clusters. The one surrounds Stellenbosch University, and the other encompasses several South African universities and research institutes. Mondi and Sappi were the companies that collaborated the most with independent research institutes. The keywords and theme analysis confirmed that the journal’s aim and scope were supported in the publications. The theme analysis also identified themes or aspects with very few publications. The methods illustrated in this paper can be used to identify research strengths and weaknesses and may assist in strategic planning for future research prioritisation.
