A case study on player selection and team formation in football with machine learning
Abstract
Machine learning has been widely used in different domains to extract information from raw data. Sports
is one of the popular domains for researchers to work on recently. Although score prediction for matches is the most
preferred application area for artificial intelligence, player selection, and team formation is also an application area worth
working on. There are some studies in the literature about player selection and team formation which are examined in
this study. The study has two important contributions: First one is to apply seven different machine learning algorithms
on our dataset to find the best player combination for the U13 team of Altınordu Football Academy and comparing the
results with that of the coach’s lineup and lineups of 20 matches played in 2019–2020 season. Second is combining the
data obtained from the trainings of the players and coach evaluations of the players and feeding the machine to make
more accurate predictions. The data from the trainings is gathered with Hit/it Assistant and the coach evaluations
of the players are stated by the golden standard according to eighteen criteria stated in the literature. Synthetically
generated data is also used in the final dataset to obtain more accurate classification results. Another remarkable aspect
of the study is that no match data is used to form the team to be proposed for the next match, instead real match
data is only used for evaluation. The results show that machine learning algorithms can be used for player selection and
team formation process because random forest algorithm, which is executed on WEKA environment, can make player
selections with 93.93% reliability and the lineup suggestions of these algorithms are 97.16% similar to coach’s ideal team
and also the best performing algorithm has an average performance of 89.36% for team formation when compared with
the match lineups of 2019–2020 football season.