Deep learning-driven decision tree ensembles for table tennis: analyzing serve strategies and first-three-stroke outcomes

This paper presents a novel artificial intelligence system that integrates deep learning-driven decision tree ensemble algorithms (DLDDTEA) for table tennis match analysis. By analyzing videos of professional matches featuring Lin Yun-Ju and Ma Long, the system extracts key insights into player techniques, hitting positions, and scoring outcomes. DLDDTEA processes the video data and constructs a predictive model to determine optimal serve positions and estimate point win/loss probabilities within the first three exchanges. The results revealed distinct serve strategies and techniques: Lin Yun-Ju favors backhands, whereas Ma Long prefers forehands. Based on these findings, this study offers specific training and strategic recommendations for both players. Thus, the proposed system offers a comprehensive framework for table tennis match analysis enabling players to gain a deeper understanding of their strengths and weaknesses, ultimately facilitating the development of more effective training and competitive strategies.