基于CPET时序聚类的中长跑耐力运动员选拔方法

心肺运动试验（CPET）能将人体的呼吸系统、心血管系统等综合为一体，不仅能够体现受试者的有氧运动能力，评估受试者的心肺耐力，而且能以整体整合医学的视角来研究受试者对运动的应激反应。为对CPET数据进行凝聚层次聚类分析，提出一种基于时间序列形态特征的算法。选取15名业余中长跑运动员的CPET数据作为聚类对象，聚类指标选取了表征有氧能力和心肺耐量的耗氧量、二氧化碳、心率、分钟通气当量、代谢当量、生理死腔与潮气量比值、呼吸商及每搏输出量等8类指标，体现运动员摄取、利用氧的效率、肺循环以及心功能等综合状况。通过聚类分析发现受试者个体差异较大，未出现明显的“群居分布”特征，根据轮廓系数评估可剔除心肺耐量较差的测试者。实验结果表明，该算法在确保聚类准确率的同时能够降低数据压缩率，且对形态特征显著的数据集进行聚类效果更佳。

Selection Method of Middle and Long Distance Endurance Athletes Based on CPET Time Series Clustering

Cardiopulmonary Exercise Testing (CPET) can integrate the respiratory and cardiovascular systems of the human body.It can reflect the aerobic exercise capacity of subjects as well as be used to evaluate their cardiorespiratory endurance and examine their stress response to exercise from the perspective of holistic integrative medicine.To perform agglomerative hierarchical clustering analysis on CPET data, we propose an algorithm based on time-series morphological features.CPET data of 15 amateur intermediate- and long-distance runners are selected as the clustering object.The following eight cluster indexes are selected to represent the aerobic capacity and cardiopulmonary tolerance:oxygen consumption, carbon dioxide, heart rate, minute ventilation equivalent, metabolic equivalent, dead volume/tidal volume, respiratory quotient, and stroke volume.These indices reflect the comprehensive status of athletes' intake and utilization of oxygen, pulmonary circulation, and cardiac function.Results of cluster analysis show that the individual differences of the subjects are insignificant, and that a clear "group distribution" feature does not exist.An evaluation of the silhouette coefficient shows that testers with low cardiorespiratory tolerance can be excluded.Experimental results show that the algorithm ensures the clustering accuracy and reduces the data compression rate, and that the clustering effect is more prominent in datasets exhibiting significant morphological characteristics.