Accurate assessment of key soil attributes such as soil organic carbon (OC), available phosphorus (P), and available potassium (K) using mid-infrared spectroscopy (MIRS) is essential for better soil management in precision agriculture. However, the calibration of the portable version of MIRS is more challenging than the benchmark technologies, hence, demanding more efficient modelling methods to provide accurate outcomes. This research aims to use the stacked generalisation machine learning (SG–ML) framework, combining support vector machine (SVM), gradient boosted regression (GBR), and random forest (RF), using linear ridge regression as a meta learner, for predicting OC, P, and K using MIR spectra of 375 soil samples collected from four farms (Flanders, Belgium). The performance of the SG–ML models was compared with the multilayer perceptron (MLP) deep learning (DL) method. Results showed the superiority of the SG–ML method over the corresponding single ML and DL models. The predictive performance of SG–ML using the validation set was excellent for the three soil attributes, with coefficient of determination (R²) and root mean square error (RMSE) values of 0.88% and 0.10%, 0.85 and 4.53 mg 100 g⁻¹, and 0.84 and 3.87 mg 100 g⁻¹ for OC, K, and P, respectively. The performance of DL models were good for OC (R² = 0.65, and RMSE = 0.17%), poor for K (R² = 0.58 and RMSE = 7.59 mg 100 g⁻¹), and very poor for P (R² = 0.46, and RMSE = 6.57 mg 100 g⁻¹). The SG–ML reduced the prediction RMSE by 10% to 31%, compared with the single ML (SVM, RF, and GBR) models. In summary, the proposed stacking method is a powerful modelling tool for the accurate prediction of key soil attributes using portable MIRS.

中文翻译：

---

将中红外光谱与堆叠泛化机器学习相结合以预测关键土壤特性

使用中红外光谱 (MIRS) 准确评估土壤有机碳 (OC)、有效磷 (P) 和有效钾 (K) 等关键土壤属性对于精准农业中更好的土壤管理至关重要。然而，便携式 MIRS 的校准比基准技术更具挑战性，因此需要更有效的建模方法来提供准确的结果。本研究旨在使用堆叠泛化机器学习 (SG–ML) 框架，结合支持向量机 (SVM)、梯度提升回归 (GBR) 和随机森林 (RF)，使用线性岭回归作为元学习器，用于预测OC、P 和 K 使用从四个农场（比利时法兰德斯）收集的 375 个土壤样品的 MIR 光谱。将 SG-ML 模型的性能与多层感知器 (MLP) 深度学习 (DL) 方法进行了比较。结果表明 SG-ML 方法优于相应的单一 ML 和 DL 模型。使用验证集的 SG-ML 对三种土壤属性的预测性能非常好，决定系数（OC、K 和 P的R ² ) 和均方根误差 (RMSE) 值分别为 0.88% 和 0.10%、0.85 和 4.53 mg 100 g ^-1以及 0.84 和 3.87 mg 100 g ^{-1 。}DL 模型的性能对于 OC（R ²  = 0.65 和 RMSE = 0.17%）很好，对于 K（R ²  = 0.58 和 RMSE = 7.59 mg·100 g ^-1）很差，对于 P（R ²  = 0.46，RMSE = 6.57 毫克 100 克^-1）。与单一 ML（SVM、RF 和 GBR）模型相比，SG–ML 将预测 RMSE 降低了 10% 至 31%。总之，所提出的叠加方法是使用便携式 MIRS 准确预测关键土壤属性的强大建模工具。