ABSTRACT
In addition to possible application in player and object tracking in digital image sequences, machine learning methods are used to extract information from a wide range of data that is collected before, during and after team games. The chapter summarizes basics and principles of machine learning and discusses approaches for analyzing different data categories. Video data form the basis for the automatic classification of game scenes, while position data is used to identify types and intensities of movement and to identify events and tactical behavior. Additional or different data categories (physiological data, fitness data, scouting data, coaching data, crowd data, etc.) are evaluated if questions about the player rating and ranking, team selection and formation, player recruitment, talent identification or injury risk assessment are to be answered. Finally, issues of prediction of the game result may be assessed.
