Inteligência computacional aplicada ao futebol americano

• Ano de defesa: 2018
• Autor(a) principal: Rodrigues, Diego da Silva
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Elétrica; UFRJ
• Palavras-chave em Português: Engenharia elétrica; Futebol americano; Modelos de ranqueamento; Ranqueamento Bayesiano; Ciência do esporte; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: http://hdl.handle.net/11422/11567

This work presents the application of computational intelligence models to support amateur level american footbal teams. Two models were developed, with the goal of supporting coaches and athletes, using data from different sources. One specialist neural network ensemble was trained, using data from a high level american team, in order to extract what game characteristics affect playcall, for a professional team. The ensemble could identify a relevant set of attributes among the ones analyzed. Using data collected for amateur level players, a ranking algorithm was developed. This model was used to rank athletes on a weekly training camp opened to all players. The model was also use privately for an amateur team from Rio de Janeiro, to rank their players during the season.