Estimativa da produtividade de trigo em função da adubação nitrogenada utilizando modelagem neuro fuzzy

Abstract

Currently new techniques for data processing, such as neural networks, fuzzy logic and hybrid systems are used to develop predictive models of complex systems and to estimate the desired parameters. In this article the use of an adaptive neuro fuzzy inference system was investigated to estimate the productivity of wheat, using a database of combination of the following treatments: five N doses (0, 50, 100, 150 and 200 kg ha(-1)), three sources (Entec, ammonium sulfate and urea), two application times of N (at sowing or at side-dressing) and two wheat cultivars (IAC 370 and E21), that were evaluated during two years in Selviria, Mato Grosso do Sul, Brazil. Through the input and output data, the system of adaptive neuro fuzzy inference learns, and then can estimate a new value of wheat yield with different N doses. The productivity prediciton error of wheat in function of five N doses, using a neuro fuzzy system, was smaller than that one obtained with a quadratic approximation. The results show that the neuro fuzzy system is a viable prediction model for estimating the wheat yield in function of N doses.