Two-stage topic modelling of scientific publications: A case study of University of Nairobi, Kenya

Abstract

Unsupervised statistical analysis of unstructured data has gained wide acceptance especially in natural language processing and text mining domains. Topic modelling with Latent Dirichlet Allocation is one such statistical tool that has been successfully applied to synthesize collections of legal, biomedical documents and journalistic topics. We applied a novel two-stage topic modelling approach and illustrated the methodology with data from a collection of published abstracts from the University of Nairobi, Kenya. In the first stage, topic modelling with Latent Dirichlet Allocation was applied to derive the per-document topic probabilities. To more succinctly present the topics, in the second stage, hierarchical clustering with Hellinger distance was applied to derive the final clusters of topics. The analysis showed that dominant research themes in the university include: HIV and malaria research, research on agricultural and veterinary services as well as cross-cutting themes in humanities and social sciences. Further, the use of hierarchical clustering in the second stage reduces the discovered latent topics to clusters of homogeneous topics.

The Hitchhiker's Guide to Integration of Social and Ethical Awareness in Precision Livestock Farming Research

“…Due to the unsupervised nature of the method which potentially could remove the human bias from the review process, and capacity to process a large number of documents at a relatively low computational cost, the use of NLP and Topic Modelling is becoming more popular in academia for explorative literature studies (Valle et al, 2014;Liu et al, 2016;Asmussen and Møller, 2019;Muchene and Safari, 2021). The interpretation of the results produced by LDA models, however, might pose a challenge if the initial hypothesis is not supported by manual overview of the text material or if the number of topics produced by the model is not cross-fold validated against the initial dataset.…”

An Application of Machine Learning Techniques to Analyze Patient Information to Improve Oral Health Outcomes

“…The extracted semantic structures are called topics and represent recurring patterns or clusters of co-occurring words in documents (27). Topics are extracted based on a probabilistic model that determines the most frequent co-occurring words over all documents (28). Key elements of TM are words or terms (a basic unit of discrete data), documents (a sequence of terms), corpus (a collection of documents), and document-term-matrix (DTM; a matrix that presents the frequency of each word in each document) (28).…”

“…Topics are extracted based on a probabilistic model that determines the most frequent co-occurring words over all documents (28). Key elements of TM are words or terms (a basic unit of discrete data), documents (a sequence of terms), corpus (a collection of documents), and document-term-matrix (DTM; a matrix that presents the frequency of each word in each document) (28). An example of a DTM is presented in Table 1 where each cell is a frequency of terms used (column) in each document (row).…”

“…Latent Dirichlet Allocation (LDA) is a common TM method and one of the most popular ML algorithms (28). LDA extracts previously unknown or latent information from an immense number of documents' original texts and unstructured data (28,29).…”

“…Latent Dirichlet Allocation (LDA) is a common TM method and one of the most popular ML algorithms (28). LDA extracts previously unknown or latent information from an immense number of documents' original texts and unstructured data (28,29). As patient charts are frequently found in the form of text files without any organized format, LDA is an appropriate method to extract salient information from these charts.…”