A Data Mining-Based Model for Evaluating Tennis Players’ Training Movements

Abstract

This paper uses data mining technology to mathematically model the training movements of tennis players, establish a three-dimensional data information database of athletes utilizing depth imaging, analyze the data with data mining algorithms, and derive the results after comparative evaluation and analysis with a database of movement characteristics of tennis dribblers. This paper uses video observation and mathematical modeling to construct a tennis player training action evaluation model, which provides a reference basis for tennis players to improve and enhance their tactical level; it can also provide a reference for the development of sports training special theory of tennis projects and enrich the tactical diagnosis method of tennis matches. To improve the accuracy of 3D human pose estimation, this paper adopts a 3D skeleton point extraction method based on RGBD images; for the action alignment problem, this paper uses a dynamic time warping (DTW) algorithm; for the similarity measure, this paper gives a Pearson correlation coefficient method based on the joint point features of human parts. This paper aims to conduct a systematic theoretical analysis of tennis players’ training movements based on theories and methods such as system science theory and social network analysis. On this basis, the characteristics of tennis training technology development are analyzed from a combination of qualitative and quantitative perspectives, while the development of tennis player training is explored based on tracking observations of tennis player movement training, and finally, the attack and service characteristics of tennis training are analyzed to better provide some reference for the sustainable development of tennis.