融合體能評估與強化學習的個性化籃球教學訓練策略
打開文本圖片集
【中圖分類號】G841 【文獻標識碼】A 【文章編號】2095-2813（2025）19-0018-05
[Abstract] With the increasing personalization of sports training demands，designing precise training programs based onathletes'diferent physicalcharacteristics hasbecomeacrucial topicinmodersports training.Basketballplayers' physical fitness levels directlyafects theeffectivenessoftheir technicalandtactical execution.Therefore，personalized training optimization is crucial for improving competitive performance.This paper proposes a personalized basketball training optimization framework based on Reinforcement Learning （RL），using physical fitnessasssment data to guide training strategyselection.Toovercomethe problem of insuficient training data，amulti-level data augmentation strategy is adopted，incuding SMOTEalgorithm，GANnetworks，and Gaussiannoiseijection techniques，toepand original samples 5-10 times. This research designs an automated training decision system based on Q-learning algorithm， which can adjust the training plan inreal time according tothe physicalfitness status of athletes.Experimental results show that the experimental group's maximum oxygen uptake has significantly increased compared to before the experiment （P<0.001） ， while mixed sample training models outperformed pure real sample models by approximately 30% This research provides new ideas and practical references for inteligentandpersonalized development in basketball training.
[Keywords]Basketball trainng;Personalizedtraining;Reinforcementlearing;itnessassessment;Dataenhancement
作者簡介：陳建樺（1997—），男，本科，研究方向為運動訓練學、體育數(shù)據(jù)分析與技術應用。（剩余8049字）