Artificial Intelligence for Tennis Performance Analysis and Match Prediction

Tennis players' technical movements, movements, spatial behavior, and other aspects during matches can already be well analyzed by artificial intelligence, but the data and game predictions obtained from these analyses still need to be integrate. This is the research direction of this narrative review. This review summarizes the relevant data sources, calculation methods, technical features, performance indicators, and prediction methods. Based on technical action information, a comprehensive discussion was conducted around the main line of "from analyzing performance indicators to final prediction", and the relationship between research at different levels was elaborated. At the same time, this review also indicates that videos, Hawk-Eye technology, and wearable devices during matches can provide sufficient information on players' movement and spatial use; These data captured through external devices and analyzed by artificial intelligence need to be combined with the competition context in order to be used as supplementary data for in-game prediction. On this basis, this review proposes research directions for connecting action analysis with whole game prediction, including unified indicators and label definitions, prediction based on standard time series, and comparison using time series validation, probability calibration, and traditional benchmark models. This comprehensive technology provides a clear analytical approach that not only helps to understand the contribution of artificial intelligence in tennis analysis, but also provides new ideas for future game predictions.