Tennis Patterns: Player, Match and Beyond

This paper proposes to apply the framework of pattern mining to the game of tennis. The method builds upon similarity measures for subrallies from a tennis match. Several mining problems are examined, where frequent subrallies (that have a high similarity to other subrallies from the same match) must be found. One can consider individual players or matches (involving rotation), and also general patterns for tennis. Experiments show some of the merits of the method.