Application of spatio-temporal data in site-specific maize yield prediction with machine learning methods

Abstract

In order to meet the requirements of sustainability and to determine yield drivers and limiting factors, it is now more likely that traditional yield modelling will be carried out using artificial intelligence (AI). The aim of this study was to predict maize yields using AI that uses spatio-temporal training data. The paper has advanced a new method of maize yield prediction, which is based on spatio-temporal data mining. To find the best solution, various models were used: counter-propagation artificial neural networks (CP-ANNs), XY-fused Querynetworks (XY-Fs), supervised Kohonen networks (SKNs), neural networks with Rectangular Linear Activations (ReLU), extreme gradient boosting (XGBoost), support-vector machine (SVM), and different subsets of the independent variables in five vegetation periods. Input variables for modelling included: soil parameters (pH, P2O5, K2O, Zn, clay content, ECa, draught force, Cone index), micro-relief averages, and meteorological parameters for the 63 treatment units in a 15.3 ha research field. The best performing method (XGBoost) reached 92.1% and 95.3% accuracy on the training and the test sets. Additionally, a novel method was introduced to treat individual units in a lattice system. The lattice-based smoothing performed an additional increase in Area under the curve (AUC) to 97.5% over the individual predictions of the XGBoost model. The models were developed using 48 different subsets of variables to determine which variables consistently contributed to prediction accuracy. By comparing the resulting models, it was shown that the best regression model was Extreme Gradient Boosting Trees, with 92.1% accuracy (on the training set). In addition, the method calculates the influence of the spatial distribution of site-specific soil fertility on maize grain yields. This paper provides a new method of spatio-temporal data analyses, taking the most important influencing factors on maize yields into account.

Spatial Variability of Soil Properties and Its Effect on Maize Yields within Field—A Case Study in Hungary

“…Zn content had a significant effect on maize yields [27]. In the examined five years, precipitation in May was identified as the most important influencing factor for maize yields in this field [22].…”

“…Zn content had a significant effect on maize yields [27]. In the examined five years, precipitation in May was identified as the most important influencing factor for maize yields in this field [22]. The pH of the soil, the clay content, the electrical conductivity of the upper layer, and the normalized difference vegetation index ranked among the top five factors.…”

“…After harvesting cereals, the conventional tillage system consists of stubble tillage (~15 cm), followed by mouldboard ploughing (~20 cm), disking, and tooth harrowing to prepare the seedbeds for winter crops. In the fall, spring cereals are ploughed, and tillage systems are prepared [22].…”

“…In summary, four types of parameters were used: (i) analysis of soil parameters over the five years: pH, P 2 O 5 , K 2 O, and Zn; measured in three years: EC a in two layers, Veris N3, and N4; measured in one year: Cone Index (MPa), pH (water), draught force (kN), (ii) crop parameters (NDVI, measured in four years), (iii) non-changing soil variables (clay content (%), relative elevation (m) evaluated in one year (2010) but used in all years). (iv) Of the 35 meteorological features observed, the following five values were measured in each month from April to October during the growing season of maize: sum of precipitation, average temperature, average relative precipitation, vaporization, evapotranspiration, and aridity index [22].…”

Stability and Adaptability of Maize Hybrids for Precision Crop Production in a Long-Term Field Experiment in Hungary

“…Usually, farmers Agronomy 2021, 11, 2167 2 of 14 continue to use a given input until the income from the consumption of each unit equals the cost of consuming it [5]. Yield prediction is an essential research topic for optimizing the use of different inputs, as it has an equally important point of reference for farm management during planning, agrotechnological intervention, and preharvest processes [6]. However, the relationship between yield and high inputs values (such as water and nitrogen fertilizer) follows the law of diminishing returns.…”