In my second post for Towards Data Science - “A Deep Dive into the Language of Football” I'm Going deeper (much deeper) into the language of football and Player2Vec, explaining football actions, player representation, and semantics. Using StatsBomb data, I aim to mitigate the explainability gap of word embeddings models both in the word (action) and the document (player) levels.
I will focus on four explainability methods that I found the most informative and reliable in practice: representation-based explainers, analogies, similarities, and creating players’ variations.
Apart from explaining predictions, using these methods, we will be able to build complex profiles for targeting players in the transfer market. For example, we can search for a player like Antoine Griezmann, but accomplish more dribbles, or like Felipe Coutinho with better play on his weaker foot.