A Behavioral System for the Detection of Injury and Rehabilitation Process using Decision Tree Model

Abstract

Imen Chebbi, Sarra Abidi and Leila Ben Ayed

To better avoid injuries in sports, prevention strategies increasingly include modern techniques like machine learning that allow for an evaluation of injury risk. This article aims to assess the injury risk for 250 athletes. The risk indicators measured daily were the athletes’ views of their physical and psychological conditions, which they self-reported each morning and evening using a customized application. The output data matched the injuries reported by the athletes. A Decision Tree model was trained and optimized to predict the incidence of an injury using the measured variables. Our model’s performance score accuracy = 99.60. Estimating the risk of injury is challenging due to the disparity between the number of injuries and observations. The prediction model identified physical and positive emotional elements as the most influential.

PDF