A Universal Dense Football Event Representation Based on TabTransformer
2026-06-08 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors developed a new way for computers to understand detailed soccer game data by using a Transformer model, which helps the computer learn the meaning behind different types of player actions instead of just treating them as simple labels. Their method converts these actions into richer codes that capture the context and relationships between actions. This helps the computer make better predictions about player performance and playing styles. Tests showed their approach predicts outcomes more accurately than older methods.
Football event dataTransformer modelSelf-attentionCategorical featuresEmbedding vectorsPlayer evaluationMatch outcome forecastingBrier scoreAction value estimationPlay style recognition
Authors
Weiran Yang, Daniel Memmert, Maximilian Klemp-Weins
Abstract
Football event data constitute a rich spatiotemporal source for quantitative analysis of player actions in team sports. These datasets contain heterogeneous features, combining continuous location coordinates with categorical variables such as action type, action outcome, and body part. Such data have been applied in sports analytics for match outcome forecasting, player evaluation, and tactical pattern recognition. However, existing approaches predominantly encode categorical features using one-hot or ordinal embedding representations, overlooking the intrinsic semantics of action descriptors. The Transformer is a deep neural network architecture based on self-attention that captures dependencies between input features at arbitrary positions. We propose and implement a Transformer-based model to learn latent dependencies among categorical event features and produce dense representations of football events. By encoding categorical features as learned embedding vectors, sport-specific action semantics are captured during pretraining, enabling the representations to support downstream tasks such as action value estimation and play style recognition. Empirical evaluation shows that the embedding representations yield superior probability calibration over task-specific baselines on the downstream prediction tasks, as measured by Brier score.