Year: 2025
Type: article
Abstract: Sports injuries can be considered as self-exciting, in that they can be linked to a previous injury or can be a new independently acquired injury. Disentangling which injuries are connected, either directly or indirectly, remains a persistent challenge, but is important for sport and athlete management. In this paper, we consider injuries as a type of Hawkes Process and view estimation as a branching process. We also propose a computationally efficient latent variable formulation in the context of clustering subsequent sporting injuries among professional Australian Football League (AFL) players in a small sample context. The results underscore the broader advantages of utilising such a point process for accurately modelling and estimating the occurrence of rare events. Our approaches are then assessed to demonstrate better estimation performance and compared to determine their suitability for injury forecasting.
Source: SN Computer Science
Authors: John Worrall, Paul Pao-Yen Wu, Liam A. Toohey, Kerrie Mengersen
Institutions: Queensland University of Technology, University of Canberra
Language: English
Subfield: Orthopedics and Sports Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $3,190
OpenAlex ID: W4409520637