Mapping knowledge structure of artificial intelligence research in Bangladesh based on co-word analysis

Abstract

Purpose: This article aims to map the knowledge structure of artificial intelligence (AI) in Bangladesh through detecting the interdisciplinarity and topic hotspots in the light of co-word analysis.

Methodology: This study adopted bibliometric analysis of publications collected from the Web of Science (WoS) database. The WoS database was searched and 1557 publications were found. 1359 papers were selected for final analysis after eliminating duplicates. Co-occurrence words matrix, keyword clusters, hot topics were mapped using co-word analysis. The results were mapped, clustered and presented by VOSviewer.
Results: The result showed a rapidly increasing publication trajectory with 12 sub-domain cluster under the AI knowledge domain in Bangladesh. It also identified that AI, machine learning, classification, neural network, deep learning, artificial neural network, convolutional neural network, support vector machine and data mining are hot topics during the period of studied time. However, the findings also suggest that many research areas in the research domain of AI of Bangladesh is still nascent.
Limitation: VOSviewer often avoid having overlapping terms when multiple terms are positioned very close to each other. So, the overlapping terms remain invisible sometimes.
Practical implications: This study may have potential usefulness in uncovering the AI research fields’ intellectual structure within a discipline and also to anticipate future innovation pathways of AI field in Bangladesh.
Originality: Bibliometric methods to explore the research trend and growth of AI research field as a ‘knowledge base’ in Bangladesh is one of the first attempts.