Classification and selection of young athletes through the analysis of cluster

Abstract

The purpose of this study were: a) to classify young athletes in three different groups (better, medium and worse), considering kinanthropometry variables. b) Compare the classification obtained between this statistical procedure and coaches’ opinion. Sample was composed by 405 young male athletes from 11 to 15 years old, divided in five age groups. Physical fitness tests and anthropometric variables were measured. In data analysis the cluster analysis was used through the k-means procedure. The coaches of the different modalities agreed in 63,5% (handball), 65,2% (soccer), 74,7% (canoeing) with the classification accomplished by the multivariate statistical model. The statistical procedure of cluster analysis can be used in sport practice, in the process of evaluation of the development of physical fitness, physical growth and young athletes' corporal composition involved in TLP.