Machine learning and the identification of Smart Specialisation thematic networks in Arctic Scandinavia

• Published in: Regional Studies
• Main Authors: Moilanen, Mikko, Østbye, Stein, Jaakko, Simonen
• Format: Article in Journal/Newspaper
• Language: English
• Published: Taylor & Francis 2021
• Subjects: VDP::Technology: 500::Information and communication technology: 550; VDP::Teknologi: 500::Informasjons- og kommunikasjonsteknologi: 550; Arctic
• Online Access: https://hdl.handle.net/10037/22909; https://doi.org/10.1080/00343404.2021.1925237

The European Union (EU) has recognized that universities and research institutes play a critical role in regional Smart Specialisation processes. Our research aims to identify thematic cross-border research domains across space and disciplines in Arctic Scandinavia. We identify potential domains using an unsupervised machine-learning technique (topic modelling). We uncover latent topics based on similarities in the vocabulary of research papers. The proposed methodology can be utilized to identify common research domains across regions and disciplines in almost real time, thereby acting as a decision support system to facilitate cooperation among knowledge producers.