Taha Zahari, Musa Rabiu Muazu, P P Abdul Majeed Anwar, Alim Muhammad Muaz, Abdullah Mohamad Razali
Innovative Manufacturing, Mechatronics and Sports Laboratory, Faculty of Manufacturing Engineering, Universiti Malaysia Pahang, 26600 Pekan, Pahang, Malaysia.
Innovative Manufacturing, Mechatronics and Sports Laboratory, Faculty of Manufacturing Engineering, Universiti Malaysia Pahang, 26600 Pekan, Pahang, Malaysia; Faculty of Applied Social Sciences, Universiti Sultan Zainal Abidin, 21300 Kuala Terengganu, Terengganu, Malaysia.
Hum Mov Sci. 2018 Feb;57:184-193. doi: 10.1016/j.humov.2017.12.008. Epub 2017 Dec 14.
Support Vector Machine (SVM) has been shown to be an effective learning algorithm for classification and prediction. However, the application of SVM for prediction and classification in specific sport has rarely been used to quantify/discriminate low and high-performance athletes. The present study classified and predicted high and low-potential archers from a set of fitness and motor ability variables trained on different SVMs kernel algorithms. 50 youth archers with the mean age and standard deviation of 17.0 ± 0.6 years drawn from various archery programmes completed a six arrows shooting score test. Standard fitness and ability measurements namely hand grip, vertical jump, standing broad jump, static balance, upper muscle strength and the core muscle strength were also recorded. Hierarchical agglomerative cluster analysis (HACA) was used to cluster the archers based on the performance variables tested. SVM models with linear, quadratic, cubic, fine RBF, medium RBF, as well as the coarse RBF kernel functions, were trained based on the measured performance variables. The HACA clustered the archers into high-potential archers (HPA) and low-potential archers (LPA), respectively. The linear, quadratic, cubic, as well as the medium RBF kernel functions models, demonstrated reasonably excellent classification accuracy of 97.5% and 2.5% error rate for the prediction of the HPA and the LPA. The findings of this investigation can be valuable to coaches and sports managers to recognise high potential athletes from a combination of the selected few measured fitness and motor ability performance variables examined which would consequently save cost, time and effort during talent identification programme.
支持向量机（SVM）已被证明是一种用于分类和预测的有效学习算法。然而，SVM在特定运动中的预测和分类应用很少被用于量化/区分高水平和低水平运动员。本研究从一组在不同SVM核算法上训练的体能和运动能力变量中，对高潜力和低潜力弓箭手进行分类和预测。从各种射箭项目中抽取的50名平均年龄为17.0±0.6岁的青年弓箭手完成了六箭射击得分测试。还记录了标准的体能和能力测量指标，即握力、垂直跳跃、立定跳远、静态平衡、上肢肌肉力量和核心肌肉力量。基于层次凝聚聚类分析（HACA），根据测试的性能变量对弓箭手进行聚类。基于测量的性能变量，训练了具有线性、二次、三次、精细径向基函数（RBF）、中等RBF以及粗糙RBF核函数的SVM模型。HACA分别将弓箭手聚类为高潜力弓箭手（HPA）和低潜力弓箭手（LPA）。线性、二次、三次以及中等RBF核函数模型在预测HPA和LPA时，表现出相当出色的分类准确率，为97.5%，错误率为2.5%。本调查结果对于教练和体育管理人员从所选择的少数几个测量的体能和运动能力表现变量中识别高潜力运动员具有重要价值，这将在人才识别计划中节省成本、时间和精力。
