Análise multivariada aplicada na construção de scores de rendimento dos principais jogadores do futebol mundial

• Autor(a) principal: Brigante, Gianpedro Robertto Mella
• Data de Publicação: 2022
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/217390
• topic: Soccer; Multivariate analysis; Performance; Principal components; Ward’s method; Futebol; Análise multivariada; Score; Rendimento; Componentes principais; Método de Ward; Análise de agrupamentos; Análise de componente principais

In this work, two methods of multivariate analysis were adopted, Principal Component Analysis and Cluster Analysis, with the aim of analyzing the performance, taking into account variables related to the attack of the main athletes of the five biggest national championships, namely: the Brasileirão Serie A, La Liga, Serie A Italia, Premier League and France Ligue. The Principal Component Analysis method was used to reduce the number of variables and simplify the players’ interpretation, in addition to providing the performance scores of each analyzed athlete. This application was very effective as it managed to extract about 84% of the information from eight correlated variables into two new uncorrelated variables. With this model it was also possible to make Biplot graphs that helped to identify the players who stood out the most in each variable due to the scores obtained. The performance of players by championship was also analyzed, allowing the comparison of these studied competitions. After obtaining the performance scores, a grouping method called Ward’s Method was used, which groups the individuals (athletes) according to their proximity according to the data, then the quality of these groups was observed by the silhouette graph that makes it possible to see if the player is well placed in his group. Making the grouping taking into account only the scores of the two components chosen, it was noticed that it was not possible to find a strong structure of the groups, but the groups were consistent with the interpretations obtained in the Biplot graphs.