Considerations about Application of Machine Learning to the Prediction of Sigatoka Disease


  • Luis Alexander Calvo-Valverde Centro de Investigaciones en Computación Instituto Tecnológico de Costa Rica DOCINADE
  • Mauricio Guzmán-Quesada Dirección de Investigaciones Corporación Bananera Nacional S.A.
  • José Antonio Guzmán-Álvarez Dirección de Investigaciones Corporación Bananera Nacional S.A.


One issue that has occupied the attention of humanity is the production of food and view it from several perspectives: the quality of the seed, the production process, diseases that affect productivity, the effect of climate and location, in example. As a contribution to the above situation, this paper presents the application of one discipline of artificial intelligence, known as machine learning, which involves the study of computer algorithms that improve automatically through experience. This type of learning has been used in applications ranging from data mining to discover rules in large datasets, to information filtering systems that automatically learn user interests. As a particular case, the Corporación Bananera Nacional of Costa Rica (Corbana) has stations measuring meteorological variables. These variables measured are temperature, precipitation, humidity, wind speed, among others. Corbana is interested in relating this information with the spread of a disease that affects production; this disease is the sigatoka. In addition, this organization recorded weekly in several of his research areas the following variables related to that disease: state of evolution, severity leaf 2, severity leaf 3, among others. With this data and using machine learning algorithms; they want to make predictions. This work presents the results of applying various machine learning algorithms to available data, in example, artificial neural network, support vector machines regression. These first conclusions will be tuned in the future.




World Conference on Computers in Agriculture, San Jose, Costa Rica, 2014