- class socceraction.vaep.VAEP(xfns=None, nb_prev_actions=3)#
An implementation of the VAEP framework.
VAEP (Valuing Actions by Estimating Probabilities) 1 defines the problem of valuing a soccer player’s contributions within a match as a binary classification problem and rates actions by estimating its effect on the short-term probablities that a team will both score and concede.
xfns (list) – List of feature transformers (see
socceraction.vaep.features) used to describe the game states. Uses
nb_prev_actions (int, default=3 # noqa: DAR103) – Number of previous actions used to decscribe the game state.
Tom Decroos, Lotte Bransen, Jan Van Haaren, and Jesse Davis. “Actions speak louder than goals: Valuing player actions in soccer.” In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pp. 1851-1861. 2019.
Transform actions to the feature-based representation of game states.
Compute the labels for each game state in the given game.
Fit the model according to the given training data.
Compute the VAEP rating for the given game states.
Evaluate the fit of the model on the given test data and labels.