Evaluating the performance of athletes in various sports using data mining and big data analytics

Huizhen Yang,Songzhen Zhang,Junpeng Zhang,Chen Wang

DOI: https://doi.org/10.1007/s00500-023-09620-9

IF: 3.732

2024-01-31

Soft Computing

Abstract:Sports effect evaluation is the main module of modern information construction. To improve the accuracy of sports effect evaluation and enhance the training level of athletes, this paper aims to conduct in-depth research on sports effect evaluation based on feature selection under the background of big data. Firstly, a sports evaluation system is designed based on the Window Presentation Foundation (WPF) that adopts a Client–Server (C/S) mode. The overall architecture of the system is proposed by expounding the functional modules and the schema of information tables in the database. To assess the impact of sports, the data mining algorithm and big data analytics technique called Random Forest (RF) is used as the core evaluation system. With the help of information gain, feature selection is performed, based on which decision trees are constructed by generating branch nodes. The random forest is created by combining all the decision trees, which helps in improving the generalization. The proposed method is experimentally verified and it is found that the overall stability of the effect evaluation system is good and it effectively improves the efficiency and accuracy of sports effect evaluation, thus enhancing the level of athletes' training. The proposed sports effect evaluation method performs excellently in the experimental validation, as seen by the remarkable F1 Score of 85.32%, average accuracy of 84.28%, precision of 86%, and recall of 82.28%. In comparison, this performs noticeably better than current approaches, proving its improved efficacy in assessing the effects of sports and confirming its ability to improve athlete training proficiency.

computer science, artificial intelligence, interdisciplinary applications