Exploring interdisciplinary nature of postgraduate research in the field of Computing using Text mining: a case study

Abstract

The aim of this article is to determine whether a compartmentalized curriculum at undergraduate lead to similar silos appearing in postgraduate research outputs. For this, data in the form of subject contents of undergraduate studies were obtained from the prospectus of various academic departments in the faculty. Additionally, data in the form of abstracts of research articles published by the postgraduate researchers in these departments were obtained. A total of 118 articles published between January 2016 and May 2020 was extracted from Scopus database. K-means algorithm was used on the corpus consisting of abstract dataset to obtain the clusters. Topic modelling using Latent Dirichlet Allocation (LDA) was then applied to obtain the main topics of the clusters. It was observed that three clusters were adequate in explaining a high percentage of variance in the data and there exists a substantial overlap in the main topics of the three clusters.