<abstract xmlns="http://www.w3.org/1999/xhtml"><p>Mathematical models that explain match outcome, based on the value of technical performance indicators (PIs), can be used to identify the most important aspects of technical performance in team field-sports. The purpose of this study was to evaluate several methodological opportunities, to enhance the accuracy of this type of modelling. Specifically, we evaluated the potential benefits of 1) modelling match outcome using an increased number of seasons and PIs compared with previous reports, 2) how to identify eras where technical performance characteristics were stable and 3) the application of a novel feature selection method. Ninety-one PIs across sixteen Australian Football (AF) League seasons were analysed. Change-point and Segmented Regression analyses were used to identify eras and they produced similar but non-identical outcomes. A feature selection ensemble method identified the most valuable 45 PIs for modelling. The use of a larger number of seasons for model development lead to improvement in the classification accuracy of the models, compared with previous studies (88.8 vs 78.9%). This study demonstrates the potential benefits of large databases when creating models of match outcome and the pitfalls of determining whether there are eras in a longitudinal database.</p></abstract>

Mathematical models that explain match outcome, based on the value of technical performance indicators (PIs), can be used to identify the most important aspects of technical performance in team field-sports. The purpose of this study was to evaluate several methodological opportunities, to enhance the accuracy of this type of modelling. Specifically, we evaluated the potential benefits of 1) modelling match outcome using an increased number of seasons and PIs compared with previous reports, 2) how to identify eras where technical performance characteristics were stable and 3) the application of a novel feature selection method. Ninety-one PIs across sixteen Australian Football (AF) League seasons were analysed. Change-point and Segmented Regression analyses were used to identify eras and they produced similar but non-identical outcomes. A feature selection ensemble method identified the most valuable 45 PIs for modelling. The use of a larger number of seasons for model development lead to improvement in the classification accuracy of the models, compared with previous studies (88.8 vs 78.9%). This study demonstrates the potential benefits of large databases when creating models of match outcome and the pitfalls of determining whether there are eras in a longitudinal database.

Modelling Match Outcome in Australian Football: Improved accuracy with large databases

La Trobe Sport and Exercise Medicine Research Centre

International Journal of Computer Science in Sport

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Mathematical models that explain match outcome, based on the value of technical performance indicators (PIs), can be used to identify the most important...

Modelling Match Outcome in Australian Football: Improved accuracy with large databases

Online veröffentlicht: 21. Aug. 2019

Seitenbereich: 80 - 92

DOI: https://doi.org/10.2478/ijcss-2019-0005

Schlüsselwörter
performance analysis, feature selection, coaching, team sports, data mining

© 2019 C. Young et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Modelling Match Outcome in Australian Football: Improved accuracy with large databases

Online veröffentlicht: 21. Aug. 2019

Seitenbereich: 80 - 92

DOI: https://doi.org/10.2478/ijcss-2019-0005

Schlüsselwörterperformance analysis, feature selection, coaching, team sports, data mining

© 2019 C. Young et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Schlüsselwörter
performance analysis, feature selection, coaching, team sports, data mining