Sistema de recomendación de jugadores de fútbol con R
Loading...
Official URL
Full text at PDC
Publication date
2025
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citation
Abstract
Este trabajo consiste en realizar una machine learnig que sea capaz de recomendar jugadores de fútbol con características similares.
El fin de este trabajo es poder crear un modelo que pueda ayudar a las direcciones deportivas de equipos más modestos a mejorar sus opciones a la hora de fichar jugadores, ya sea para mejorar posiciones que necesiten, o para adelantarse a cubrir las bajas de sus jugadores en plantilla.
Para su estudio se ha utilizado una base de datos con estadísticas de los jugadores que han participado al menos un partido en cualquier club de las cinco grandes ligas europeas según el ranking de la UEFA (UEFA, 2025), las cuales son:
• LaLiga.
• Premier League.
• Bundesliga.
• Serie A.
• Ligue 1.
Los datos han sido tomados de la web de estadísticas de fútbol fbref (Fbref, 2025) a través del uso de la biblioteca worldfootballR (JaseZiv, 2020) en RStudio.
Para su realización se han usado distintas técnicas de depuración de datos, análisis cluster, modelos random forests y medidas de distancias euclídeas. Estás técnicas han sido empleadas en R, debido a que es un sistema de libre acceso, y que tiene un interfaz muy fácil para su entendimiento lo que podría ayudar a su implementación en clubes con menos margen presupuestario.
Abstract: This project involves the development of a machine learning model capable of recommending football players with similar statistical profiles. The main goal is to create a tool that can support the scouting departments of local or regional clubs in improving their recruitment strategies- either by identifying suitable players to strengthen specific positions or by anticipating potential departures. To achieve this, a dataset was compiled containing statistics of players who have participated in at least one match in any club within the top European leagues, as ranked by UEFA (UEFA, 2025). These leagues are: • LaLiga. • Premier League. • Bundesliga. • Serie A. • Ligue 1. The data was sourced from the football statistics website fbref (Fbref, 2025) using the worldfootballR package (JaseZiv, 2020) in RStudio. The methodology involved data cleaning techniques, cluster analysis, random forest models, and Euclidean distance metrics. These methods were applied using R, an open-source programming environment with a user-friendly interface, making it suitable for implementation by clubs with limited budgets.
Abstract: This project involves the development of a machine learning model capable of recommending football players with similar statistical profiles. The main goal is to create a tool that can support the scouting departments of local or regional clubs in improving their recruitment strategies- either by identifying suitable players to strengthen specific positions or by anticipating potential departures. To achieve this, a dataset was compiled containing statistics of players who have participated in at least one match in any club within the top European leagues, as ranked by UEFA (UEFA, 2025). These leagues are: • LaLiga. • Premier League. • Bundesliga. • Serie A. • Ligue 1. The data was sourced from the football statistics website fbref (Fbref, 2025) using the worldfootballR package (JaseZiv, 2020) in RStudio. The methodology involved data cleaning techniques, cluster analysis, random forest models, and Euclidean distance metrics. These methods were applied using R, an open-source programming environment with a user-friendly interface, making it suitable for implementation by clubs with limited budgets.







