ABSTRACT
This chapter tackles the problem of modelling sports with the purpose of estimating the characteristics that lead an athlete to succeed or not and a team to lose or win a match. For that purpose, it takes into consideration the contributions presented in the previous chapter to move from classification, through explanation, to prediction, with the intent to estimate various parameters that highly influence the functioning of a performance environment, including athlete’s position, actions, and muscle fatigue. This then leads to propose a novel sport prediction model, through which deep learning is used as a learning strategy. The particular case of predicting the number of goal attempts and goals scored is described.
