Research Article
Application of Latent Dirichlet Allocation Topic Model in Identifying 4IR Research Trends
@INPROCEEDINGS{10.1007/978-3-031-35883-8_6, author={Muthoni Masinde}, title={Application of Latent Dirichlet Allocation Topic Model in Identifying 4IR Research Trends}, proceedings={Emerging Technologies for Developing Countries. 5th EAI International Conference, AFRICATEK 2022, Bloemfontein, South Africa, December 5-7, 2022, Proceedings}, proceedings_a={AFRICATEK}, year={2023}, month={7}, keywords={Latent Dirichlet Allocation (LDA) Fourth Industrial Revolution (4IR) Technology Trends Bibliometric Analysis and Topic Models}, doi={10.1007/978-3-031-35883-8_6} }- Muthoni Masinde
Year: 2023
Application of Latent Dirichlet Allocation Topic Model in Identifying 4IR Research Trends
AFRICATEK
Springer
DOI: 10.1007/978-3-031-35883-8_6
Abstract
The dynamic nature of the technologies associated with the fourth Industrial Revolution (4IR) presents complex scenarios for researchers, practitioners and policymakers alike. To this end, reaching decisions such as what technology to invest/train in could be made easier through a 4IR technology trend predictive tool. In this paper, we apply Latent Dirichlet Allocation (LDA) topic model to identify and predict trends in 4IR technologies. The LDA models were developed and trained using text composed of abstracts, titles and keywords retrieved from 11,7314-IR related to the 2012 to 2022 publications in the Web of Science database. The effectiveness of the resulting tool was then evaluated using text from email message distributed to subscribers of the IEEE’s Tccc-announce mailing list. From the results, our model correctly identifies trends in the following 4IR technologies and applications domains: Internet of Things, Artificial Intelligence/Machine Learning, Big Data/Data Analytics, Augmented Reality, Smart Manufacturing, Supply Chains, Sustainability and Circular Economy. By plotting and visualizing these trends over time (2019 to 2022), the validation text confirms our tool’s ability to identify the trajectory developments as identified by other similar tools such as Bibliometric Analysis.
