Identifying playing talent in professional football using artificial neural networks

Tools

Barron, D., Ball, G. ORCID: 0000-0001-5828-7129, Robins, M. and Sunderland, C. ORCID: 0000-0001-7484-1345, 2020. Identifying playing talent in professional football using artificial neural networks. Journal of Sports Sciences, 38 (11-12), pp. 1211-1220. ISSN 0264-0414

Preview

Text
14896_Sunderland.pdf - Post-print
Download (2MB) | Preview

Official URL: https://doi.org/10.1080/02640414.2019.1708036

Abstract

The aim of the current study was to objectively identify position-specific key performance indicators in professional football that predict out-field players league status. The sample consisted of 966 out-field players from 2 seasons in the Football League Championship. Players were assigned to one of three categories (0, 1 and 2) based on where they completed most of their match time in the following season, and then split based into five positions. 340 performance, biographical and esteem variables were analysed using a Stepwise Artificial Neural Network approach. A Monte Carlo cross-validation procedure was used to avoid over-fitting and the neural network modelling involved a multi-layer perceptron architecture with a feed-forward backpropagation algorithm. The models correctly predicted between 72.7% and 100% of test cases (Mean prediction of models = 85.9%), the test error ranged from 1.0% to 9.8% (Mean test error of models = 6.3%). Variables related to passing, shooting, regaining possession and international appearances were key factors in the predictive models. This is highly significant as objective position-specific predictors of players league status could be used to aid the identification and comparison of transfer targets as part of the due diligence process in professional football.

Item Type:

Journal article

Publication Title:

Journal of Sports Sciences

Creators:

Barron, D., Ball, G., Robins, M. and Sunderland, C.

Publisher:

Routledge

Date:

2020

Volume:

Number: