Self-Organising and Self-Learning Model for Soybean Yield Prediction

Abstract

Machine learning has arisen with advanced data analytics. Many factors influence crop yield, such as soil, amount of water, climate, and genotype. Determining factors that significantly influence yield prediction and identify the most appropriate predictive methods are important in yield management. It is critical to consider and study the combination of different crop factors and their impact on the yield. The objectives of this paper are: (1) to use advanced data analytic techniques to precisely predict the soybean crop yields, (2) to identify the most influential features that impact soybean predictions, (3) to illustrate the ability of Fuzzy Rule-Based (FRB) subsystems , which are self-organizing, self-learning, and data-driven, by using the recently developed Autonomous Learning Multiple-Model First-order (ALMMo-1) system, and (4) to compare the performance with other well-known methods. The ALMMo-1 system is a transparent model, which stakeholders can easily read and interpret. The model is a datadriven and composed of prototypes selected from the actual data. Many factors affect the yield, and data clouds can be formed in the feature/data space based on the data density. The data cloud is the key to the IF part of FRB subsystems , while the THEN part (the consequences of the IF condition) illustrates the yield prediction in the form of a linear regression model, which consists of the yield features or factors. In addition, the model can determine the most influential features of the yield prediction online. The model shows an excellent prediction accuracy with a Root Mean Square Error (RMSE) of 0.0883, and Non-Dimensional Error Index (NDEI) of 0.0611, which is competitive with state-of-the-art methods.

Machine learning algorithms for soybean yield forecasting in the Brazilian Cerrado

“…9 ML algorithms such as random forest, neural networks and support vector machine were successfully used to forecasting crop yield. 10 Studies such as Cai et al 11 integrated several data sources to forecasting wheat production in Australia from 2000 to 2014; the authors used the LASSO regression method and three main methods of ML (support vector machine (SVM), random forest (RF) and neural network) to build various empirical models for yield forecasting, and confirmed that the combination of climate and satellite data could achieve high forecast performance with an R 2 of 0.75.…”

Modeling of tobacco loosening and conditioning system using LSTM recurrent neural networks

“…Data-driven modeling can use a large amount of data from factory production history to train neural networks, so as establishing the statistical relationship between output variables and input variables. Fuzzy inference system [8], [9], support vector regression [10] and neural network [10], [11], can available to establish data-driven models. Due to the advantages of fuzzy system and neural network, namely comprehensiveness, transparency, adaptability and rapid convergence, fuzzy neural network has been successfully applied to solve the modeling problems in practical industrial engineering [12].…”

Identificação de sistema não linear multivariável com modelos fuzzy ALMMo de primeira ordem

“…Diversos trabalhos nas áreas de sistemas inteligentes evolutivos e sistemas de aprendizagem autônoma vêm sendo realizados na literatura, como por exemplo em Alghamdi et al (2019), em que é feita uma previsão precisa dos rendimentos da safra de soja em plantações, através de técnicas de análise de dados avançada em conjunto com modelos autônomos evolutivos; em Lughofer et al (2017), em que é feita uma compensação autônoma de desvios de dados para sistemas fuzzy evolutivos generalizados, através da divisão de regra incremental; e em Pratama et al (2018), em que é proposto um classificador de conjuntos fuzzy evolutivo, chamado de Parsimonious Ensemble (pENsemble). A aplicação da lógica fuzzy na identificação e controle de sistemas não lineares também tem sido bastante pesquisada na literatura, como por exemplo em Zhang e Shin (2021), em que um modelo fuzzy Takagi-Sugeno é construído a partir de dados experimentais de um sistema não linear desconhecido (analiticamente), permitindo a implementação de controladores ótimos para o mesmo; em Farid et al (2017), em que sistemas fuzzy do tipo 2 são utilizados para lidar com incertezas em uma planta dinâmica, aplicandose na identificação e controle adaptativo do sistema; e em Bulatov e Kryukov (2018), em que um sistema neuro-fuzzy é utilizado para identificar os modos de operação de uma usina de geração distribuída e fazer o controle adaptativo da sintonia dos reguladores automáticos de excitação (AER) e dos reguladores automáticos de velocidade do rotor (ARRS) da usina.…”