Soccer’s AI transformation: deep learning’s analysis of soccer’s pandemic research evolution

Abstract

Introduction: This paper aims to identify and compare changes in trends and research interests in soccer articles from before and during the COVID-19 pandemic. Methods: We compared research interests and trends in soccer-related journal articles published before COVID-19 (2018–2020) and during the COVID-19 pandemic (2021–2022) using Bidirectional Encoder Representations from Transformers (BERT) topic modeling. Results: In both periods, we categorized the social sciences into psychology, sociology, business, and technology, with some interdisciplinary research topics identified, and we identified changes during the COVID-19 pandemic period, including a new approach to home advantage. Furthermore, Sports science and sports medicine had a vast array of subject areas and topics, but some similar themes emerged in both periods and found changes before and during COVID-19. These changes can be broadly categorized into (a) Social Sciences and Technology; (b) Performance training approaches; (c) injury part of body. With training topics being more prominent than match performance during the pandemic; and changes within injuries, with the lower limbs becoming more prominent than the head during the pandemic. Conclusion: Now that the pandemic has ended, soccer environments and routines have returned to pre-pandemic levels, but the environment that have changed during the pandemic provide an opportunity for researchers and practitioners in the field of soccer to detect post-pandemic changes and identify trends and future directions for research. Copyright © 2023 Lee, Song, Kim, Park and Han.