Uniform Manifold Approximation and Projection Analysis of Soccer Players

In professional soccer, the choices made in forming a team lineup are crucial for achieving good results. Players characterized by different skills and their relevance depends on position that they occupy pitch. Experts can recognize similarities between players styles, but procedures adopted often subjective prone to misclassification. The automatic recognition of players’ styles based diversity help coaches technical directors prepare competition, substitute injured during season, or hire fill gaps created teammates leave. paper adopts dimensionality reduction, clustering computer visualization tools compare soccer set attributes. numerical vectors embedding particular these objects then compared means suitable distances. intermediate data is processed generate meaningful representations original dataset according (dis)similarities objects. results show adoption processing complex datasets key modeling option with current computational resources.