Embedding ML/Dl Applications in the Ecosystem of Mass Participation Tennis

Titelangaben

Buck, Christoph ; Gutheil, Niklas ; Ifland, Sebastian:
Embedding ML/Dl Applications in the Ecosystem of Mass Participation Tennis.
2023
Veranstaltung: 31st European Sport Management Conference 2023 , Sep 12 - 15, 2023 , Belfast City, North Ireland, GB.
(Veranstaltungsbeitrag: Kongress/Konferenz/Symposium/Tagung , Paper )

Angaben zu Projekten

Abstract

Machine learning (ML) and deep learning (DL) applications offer numerous ways of supporting players, coaches, teams, and other participants in sports and can co-create value for the entire system. They can be used to prevent injuries, identify talents, develop tactics, and many more. Amongst other sports, mass participation in tennis shows various operation areas for those applications, like supporting the relationship between players and coaches. Therefore, integrating ML/DL applications into the system can counteract problems like the lack of coaches or people playing with poor technique through their supporting role.

However, most endeavors in research do not reach practice since their thoughtful embedding in the system of mass participation in tennis fails. Therefore, we pose the question: How can ML/DL-based applications be embedded in mass participation tennis to co-create value in the system? We aim to answer this question by developing a framework that shows the relevant stakeholders and their relationships with the ML/DL application.

Much technology-oriented research shows that ML/DL applications can support sports stakeholders in multiple ways. Through the rapid development of technologies, computer vision (CV) technology can take over tasks that coaches currently perform. Existing artifacts enable ball tracking, injury prevention, stroke analysis, and more. It shows that ML/DL applications in sports can counteract the lack of coaches and tackle the problem of people playing with poor technique.

Despite vast research on optimizing the speed and accuracy of those ML/DL artifacts, the embedding in the system and the impact on value co-creation and stakeholder relations remain unexplored. Designing embedded applications is crucial for leveraging the possibilities of the technological approaches by nurturing its acceptance and, consequently, enabling its potential for value creation within the system. Therefore, we identify a gap in research dealing with embedding ML/DL applications in sports, which we address through mass participation in tennis.

Before dealing with the problem of embedding ML/DL applications in tennis, we developed an ML/DL human activity recognition (HAR) prototype trained to analyze a tennis player's stroke movements and provide real-time warnings and suggestions for improving stroke movement and playing technique. The high quality of the analysis and recommendation we drew from the prototype proved the feasibility of ML/DL applications to create value for the stakeholders of tennis sport. Following the guidelines and research design of Hevner (2007), we went on by identifying relevant stakeholders and their relations and, by drawing on the existing work of Raisch and Krakowski (2021), Diel et al. (2021), and Woratschek et al. (2014), built a framework for embedding ML/DL applications in the system of mass participation in tennis. After integrating justificatory knowledge from reviewing existing literature and integrating the researchers' experience and knowledge, we synthesize the framework from the foundations. In this step, we draw on the sport value framework (SVF) to explain how stakeholders and applications can co-create value (Woratschek et al. 2014). We use the automation-augmentation paradox to propose mindful handling of applications regarding coaches, who fear being replaced and might counteract the implementation (Raisch and Krakowski 2021). We use the differentiation between on-field applications and off-field technology to describe the ML/DL application (Diel et al. 2021). We then evaluate and further develop the framework by interviewing experts from different areas. In this step, we use the developed ML/DL HAR prototype to explain and evaluate the framework with 12 experts. Based on the framework, we discuss the impact of ML/DL applications on value co-creation for all affected stakeholders in tennis and transfer the results to other sports.

The framework explains the proposed relationships between the stakeholders and the ML/DL application. The developed framework shows how the different entities interact with each other to co-create value on different layers using the ML/DL application as connecting layer. The framework includes the relevant stakeholders in mass participation tennis and shows their relationship between each other and ML/DL applications. Based on the framework, we discuss potentials for co-creating value on the intra-, micro- and meso-level (Woratschek et al. 2014) and discuss changes for the stakeholders after embedding the ML/DL application.

By proposing a way of embedding ML/DL applications in sports, the framework helps technology-oriented researchers to put their artifacts in the contextual environment. Further, it guides affected stakeholders how to deal with the rise of technology in their field of expertise. Through discussing the impact of ML/DL applications on value co-creation in sports, we contribute to the research community and practitioners in sport management and information technology by developing guidelines to successfully implement technology-based artifacts in the system of mass participation tennis and comparable sports.