<b>Estimating soybean yields with artificial neural networks

Abstract

ABSTRACT. The complexity of the statistical models used to estimate the productivity of many crops, including soybeans, restricts the use of this practice, but an alternative is the use of artificial neural networks (ANNs). This study aimed to estimate soybean productivity based on growth habit, sowing density and agronomic characteristics using an ANN multilayer perceptron (MLP). Agronomic data from experiments conducted during the 2013/2014 soybean harvest in Anápolis, Goiás State, B razil, were used to conduct this study after being normalized to an ANN-compatible range. Then, several ANNs were trained to choose the best-performing one. After training the network, a performance analysis was conducted to select the ANN with a performance most appropriate for the problem, and the selected network had a 98% success rate with training data and a 72% data validation accuracy. The application of the MLP to the data used in the experiment shows that it is possible to estimate soybean productivity based on agronomic characteristics, growth habit and population density through AI. Estimativa da produtividade de soja com redes neurais artificiais RESUMO. Para estimar a produtividade de muitas culturas, incluído a soja, são utilizados modelos estatísticos complexos, que torna restrito o acesso a essa prática. Uma alternativa a estes modelos é a utilização de sistemas computacionais empregando Redes Neurais Artificiais (RNA). Este trabalho teve por objetivo estimar a produtividade da soja baseada nos hábitos de crescimento, densidade de semeadura e características agronômicas usando RNA do tipo Multilayer Perceptron (MLP). Para isto foram utilizados dados agronômicos da cultura da soja obtidos em experimento conduzido na safra 2013/2014 em Anápolis, Estado de Goiás, Brasil, cujos dados foram normalizados em intervalo compatível para trabalho com RNA e em seguida feito o treinamento de várias RNAs para a escolha da RNA com melhor performance. Após o treinamento das redes, foi realizada a análise de performance para seleção da RNA com a performance mais adequada ao problema. A RNA selecionada apresentou um índice de acerto de 98% com os dados do treinamento e um acerto de 72% com dados de validação. A aplicação das RNAs do tipo MLP nos dados do experimento conduzido demonstram que é possível estimar a produtividade da soja baseando-se nas características agronômicas, hábito de crescimento e densidade populacional por meio da IA. Palavras-chave:Glycine max (L.) Merrill; características agronômicas; modelagem; MLP; perceptron.

Estimation and forecasting of soybean yield using artificial neural networks

“…This disadvantage can be overcome using ANNs that can perform nonlinear modeling without prior knowledge of the relationships between input and output variables, so ANNs are a tool for more general and flexible modeling with respect to forecasting (Vikas & Dhaka, 2014). Beyond that, Alves et al (2018) affirm that the estimation of crop yields ANNs is attractive because it is simple and has commonly high accuracy. According to Oikawa and Ishiki (2013), ANNs can be defined as a statistical tool whose principle of operation is governed by a mathematical model inspired by F I G U R E 2 Representation of data in the input layer for 1 yr. Legend: T-Air temperature (˚C); P-precipitation (mm); CET-crop evapotranspiration (mm); AET-actual evapotranspiration of crop (mm); STO-storage (mm); DEF-deficit (mm); EXC-surplus (mm); S-sowing; V-vegetative phase; R-reproductive phase the functioning of the intelligent organisms, which acquire knowledge through experience and by processing information, and generating an output (predicted data) from one or more inputs presented (predictors).…”

“…Alves et al. (2018) used the Multilayer Perceptron (MLP) Artificial Neural Network (ANN) to estimate soybean [ Glycine max (L.) Merr.] yield in Anápolis, in the state of Goiás, Brazil.…”

Artificial Intelligence Simulating Grain Productivity During the Wheat Development Considering Biological And Environmental Indicators

“…On this perspective, detaches the studies conducted with the optimization of oats seeding rate with grain yield forecast (Dornelles et al, 2018), eucalypt volume simulation considering biological different parameters of the specie (Bhering et al, 2015), tree height simulation with different growth conditions (Campos et al, 2016). Soybean yield simulation with agronomic features, growth habits and population density (Alves et al, 2018). In maize, Soares et al (2015) had used morphological variable to simulate the crop grain yield using multilayer ANN with backpropagation training algorithm.…”

Mineral supplementation of soybean seeds with different initial nutrient levels

“…Soybean [Glycine max (L.) Merrill] is the main legume used as the raw material for production of more than half of all vegetable oils and provides approximately two thirds of the protein consumed in the world (Divito, Echeverríaet, Andrade, & Sadras, 2015;Alves et al, 2018). Thus, use of high-level technology together with management techniques is necessary to guarantee satisfactory establishment and yield increases (Spiertz & Ewert, 2009).…”