Presentation Title

Time to hamstring injury in soccer players

Faculty Mentor

Valerie Poynor

Start Date

17-11-2018 12:30 PM

End Date

17-11-2018 2:30 PM

Location

HARBESON 61

Session

POSTER 2

Type of Presentation

Poster

Subject Area

physical_mathematical_sciences

Abstract

In sports analytics, researchers are interested in modeling the time to injury of players in order to establish prevention practices and improve player longevity. The mean residual life function offers additional insight into this setting. The mean residual life function describes the expected time to injury given the player is without injury at a particular time, t. If the function is increasing this implies the player is less likely to be injured as time goes on, whereas a decreasing shape implies the player is more likely to be injured over time. In our research, we consider an artificial dataset arising from a population extracted from a real dataset. The artificial dataset describes the time to hamstring strain injury for 100 soccer players. Several of the players had injury times that were right censored, meaning that they had not experience an injury by the last recorded time in the study. The data also consists of several covariates for each player such as weight, height, and score associated with a new test procedure. Particular interest lies in assessing the test score as a risk factor for injury. We apply a number of parametric survival regression models to these data and select our final model via k-folds cross-validation. We identify the covariates that are significant risk factors for time to injury as well as provide inferential results for the mean residual life function.

This document is currently not available here.

Share

COinS
 
Nov 17th, 12:30 PM Nov 17th, 2:30 PM

Time to hamstring injury in soccer players

HARBESON 61

In sports analytics, researchers are interested in modeling the time to injury of players in order to establish prevention practices and improve player longevity. The mean residual life function offers additional insight into this setting. The mean residual life function describes the expected time to injury given the player is without injury at a particular time, t. If the function is increasing this implies the player is less likely to be injured as time goes on, whereas a decreasing shape implies the player is more likely to be injured over time. In our research, we consider an artificial dataset arising from a population extracted from a real dataset. The artificial dataset describes the time to hamstring strain injury for 100 soccer players. Several of the players had injury times that were right censored, meaning that they had not experience an injury by the last recorded time in the study. The data also consists of several covariates for each player such as weight, height, and score associated with a new test procedure. Particular interest lies in assessing the test score as a risk factor for injury. We apply a number of parametric survival regression models to these data and select our final model via k-folds cross-validation. We identify the covariates that are significant risk factors for time to injury as well as provide inferential results for the mean residual life function.