Technical knowledge extraction applied to modeling of occurrence (Cercospora coffeicola Berkeley & Cooke) coffee in the southern region of Minas Gerais

• Autor(a) principal: Souza, Vanessa Cristina Oliveira de
• Data de Publicação: 2013
• Outros Autores: Cunha, Rodrigo Luz da, Andrade, Livia Naiara, Volpato, Margarete Marin Lordelo, Carvalho, Vicente Luiz de, Esmin, Ahmed Ali Abdalla
• Tipo de documento: Artigo
• Idioma: eng; por
• Título da fonte: Coffee Science (Online)
• Texto Completo: https://coffeescience.ufla.br/index.php/Coffeescience/article/view/366
• topic: Coffea arabica L.; epidemiological data mining; epidemiology of plant diseases; Coffea arabica; mineração de dados epidemiológicos; epidemiologia de doenças de plantas

The survey of the progress of Cercospora leaf spot becomes potentially useful and understandable in understanding the disease process and in decision making for control measures. In the last years, computer programs have helped to elucidate what factors are biotic or abiotic more representative. The aim of this work was to investigate, using knowledge extraction techniques, which phenological and environmental attributes most influence on the occurrence of Cercospora leaf spot on coffee trees in southern Minas Gerais, under two tillage systems: conventional and organic. For this, data were organized incidence of Cercospora leaf spot in both cropping systems, with climatic data and phenological crop in a period of five years of evaluation. Then an algorithm based on knowledge extraction decision tree was used to obtain the attributes that most favor the occurrence of Cercospora leaf spot. The generated models were 60% hit rate and showed that the average temperature of the attribute was greater influence on the entire data and the conventional culture system. In organic management, the precipitation and phenology are the factors that most influence the occurrence of disease.