Forecasting injuries in sport
Download as PPTX, PDF
0 likes191 views
How to use data and trees to measure injury risk in sport Presented at second Data Beers Tuscany by Alessio Rossi
![[] any illness related to training load are
commonly viewed as preventable
Gabbett, 2016
In letteratura come interpretano il problema?
Analisi monodimensionali
Gabbett 2016 and Rogalski B et al. 2013](https://image.slidesharecdn.com/databeersalessiorossi-170121233627/85/Forecasting-injuries-in-sport-4-320.jpg)
Forecasting injuries in sport
- 1. ROSSI ALESSIO alessio.rossi@unimi.it Phone +39 3472339050 PREVEDERE GLI INFORTUNI SPORTIVI CON I DATI
- 2. E possibile prevedere gli infortuni dai dati di allenamento?
- 3. 24,360 Giorni di assenza 16.23% of season absence 188,058,072 Alto costo
- 4. [] any illness related to training load are commonly viewed as preventable Gabbett, 2016 In letteratura come interpretano il problema? Analisi monodimensionali Gabbett 2016 and Rogalski B et al. 2013
- 5. Il problema di predizione degli infortuni lo interpretiamo come classification problem
- 6. Features Features di allenamento (GPS) Total Distance High Speed Running (>19.8 km/h) Metabolic Distance (>20W/kg) High Metabolic Load Distance (>25.5 W/Kg) High Metabolic Load Distance Per Minute Explosive Distance (>25 W/kg <19.8 Km/h) Accelerations >2m/s2 Accelerations >3m/s2 Decelerations >2m/s2 Decelerations >3m/s2 Dynamic Stress Load (>2g) Fatigue Index (Dynamic Stress Load/Speed Intensity) Features dei calciatori Age Height Weight Role Previous injuries Injury Numero di infortuni che un giocatore ha avuto fino a quel momento. 0 = no-infortunio 1 = infortunato
- 7. Features Features di allenamento (GPS) Total Distance High Speed Running (>19.8 km/h) Metabolic Distance (>20W/kg) High Metabolic Load Distance (>25.5 W/Kg) High Metabolic Load Distance Per Minute Explosive Distance (>25 W/kg <19.8 Km/h) Accelerations >2m/s2 Accelerations >3m/s2 Decelerations >2m/s2 Decelerations >3m/s2 Dynamic Stress Load (>2g) Fatigue Index (Dynamic Stress Load/Speed Intensity) Features dei calciatori Age Height Weight Role Previous injuries Injury Numero di infortuni che un giocatore ha avuto fino a quel momento. 0 = no-infortunio 1 = infortunato
- 10. Decision Tree
- 11. Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 7 Previous injuries 1.36 1.36 1.08 2.05 1.08 1.51 0.40 Acceleration above 2 m揃s-2 53.96 57.93 49.20 57.96 78.42 77.16 --- HML per min --- 7.53 6.80 --- --- --- --- Acceleration above 3 m揃s-2 --- --- --- 16.67 --- --- --- Probability of injury 80% 50% 33% 33% 12% 9% 6% Number of injury observed 3 1 1 2 1 2 11 Number of injury predicted 1 1 1 1 0 1 8 Injury scenarios
- 12. Potentially avoided injuries 43%
- 13. Take home message Data mining 竪 utile per la predizione degli infortuni In questo dataset la variabile Previous injuries influenza gli infortuni successivi In uno scenario teorico noi otteniamo F1-score = 0.78 In uno scenario reale noi otteniamo un massimo di F1-score = 0.40
- 14. ROSSI ALESSIO alessio.rossi@unimi.it Phone +39 3472339050 GRAZIE DELLATTENZIONE