Artificial neural networks in modeling of environmental time series for yerba-mate growth dynamics

  • Fabio Takeshi Matsunaga Universidade Estadual de Londrina, Paraná, UEL
  • Miroslava Rakocevic Instituto Agronómico do Paraná - IAPAR
  • Jacques Duílio Brancher Universidade Estadual de Londrina, Paraná
Keywords: backpropagation algorithm, multilayer perceptron, shoot elongation, temperature


The artificial neural networks (ANN) are a solution to model the nonlinear systems. The monthly mean values of environmental and morphological data were used to build time series related to yerba-mate growth. Time series and data relative to rhythmic growth were used for ANN training. The final results of this FSPM are yerba-mate mock-ups, related to two particular environmental conditions.

4A Structural development of plants and light environment