Uma técnica de aprendizagem de máquina baseada em árvores de decisão: um estudo de caso em processos avícolas

Abstract

The objective of this study is to develop a computational prototype to support poultry farmers in poultry management activities. The poultry management comprises the tasks performed in the care of birds, in this case, chicken. It is a complex activity involving many factors, such as: the control of the amount of feed consumed by the birds, water, light, humidity, and other housing conditions. For the development of this prototype it was conducted a study with an expert in the poultryfield management, to understand the main activities and analyze the needs of this domain, as well as studies about the development of this production in Brazil and the functioning of the aviaries. For the development and implementation of this prototype it was still used, studies previously conducted, machine learning techniques and algorithms of decision trees, technologies that can be used to emerge a decision support system in this poultry process management.