Modelagem para detecção de talentos desportivos em jodocas.

AUTOR(ES)
DATA DE PUBLICAÇÃO

2004

RESUMO

Our objective was to create statistical models to avaluate the classification probability in the judo players of the juvenile category in the Championships of Pernambucano (local), North-Northeast (regional), Brazilian (national) and South American (international). The data were supplied by the Center of Sporting Excellency (CENESP) located in the University of Pernambuco, Recife Brazil. A total of 42 male judo players from Judos Federation of Pernambuco took part in this research. For the development of the models, the variables were: corporal mass, stature, span, dynamometric, agility, vertical impulse, located abdominal resistance, test of bar, wingate-weight, wingate-percentile of fatigue, time of practice, training phase and the information if athletes have been or not champions in one of those championships: Pernambuco, North-Northeast, Brazilian and South American. The athletes classification was related with the variables through the model of regression lineal multiple. To select the significant variables (P<0,05), the process of Stepwise, associate to the transformation of Box and Cox was used. The test of Shapiro-Wilks was applied to verify the normality of the data. The models were appraised also for ANOVA, residue analysis and the test of Durbin-Watson (D). The significant variables for the classification in the Championship of Pernambuco and North-Northeast was the span, the training phase, the time of practice and the flexibility, with a R2=98,37% and D=1,90. At the national and international level ( Championships Brazilian and South American), they were made presents the time of practice, the bar test and the training phase, presented a R2 =97,29% and D=1,40.

ASSUNTO(S)

judocas ciências agrárias talento desportivo biometria judo players sport talent modelagem modelling

ACESSO AO ARTIGO

Documentos Relacionados