发表机构
Osnabrück University; Leibniz Institute for Agricultural Engineering and Bioeconomy (ATB); German Research Center for Artificial Intelligence (DFKI)（奥斯纳布吕克大学; 莱布尼茨农业工程与生物经济研究所; 德国人工智能研究中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对土壤光谱预测难题，对比多种模型与降维方法，发现结合偏最小二乘潜在变量的表格基础模型TabPFN在田间及大规模光谱库任务中表现最优，为光谱校准模型选择提供了循证指导。
AI 中文摘要
可见光-近红外（vis-NIR）与中红外（MIR）光谱技术可实现土壤属性的快速、低成本预测。然而，将高维、高度共线性的光谱转化为准确的土壤属性预测值仍具挑战性，尤其在使用机器学习方法时。我们系统研究了土壤计量学开放基准数据集中85项回归任务的光谱建模所用回归模型与降维方法，涵盖田间尺度数字土壤制图及全球土壤光谱库场景。我们对比了上下文学习型表格基础模型（TabPFN）、卷积神经网络（CNN）、基于规则的回归模型（Cubist）、随机森林（Random Forest），以及使用全光谱、主成分分析（PCA）提取的特征、偏最小二乘（PLS）潜在变量的偏最小二乘回归（PLSR）。TabPFN在各尺度任务中始终表现出最佳整体性能，包括涉及数万个土壤样本的大规模光谱库任务。值得注意的是，直接将TabPFN应用于全光谱已优于所有经典基准模型，表明无需严格进行显式降维即可实现良好性能。通过PLS进一步提升了性能，PLS被证实是适用于所有模型的有效降维策略。将PLS潜在变量与TabPFN结合，得到了整体最佳预测结果。本研究为不同操作尺度下的光谱校准模型选择提供了循证指导，证明了PLSR的长期优势与现代表格基础模型在化学计量学中可形成互补。
英文摘要
Visible and near-infrared (vis-NIR) and mid-infrared (MIR) spectroscopy enable rapid, cost-effective prediction of soil properties. Yet, translating high-dimensional, highly collinear spectra into accurate soil property predictions remains challenging, particularly when employing machine learning. We systematically investigated regression models and dimensionality reduction approaches for spectroscopic modeling across 85 regression tasks from open benchmark datasets in pedometrics spanning field-scale digital soil mapping and a global soil spectral library. We compared an in-context learning tabular foundation model (TabPFN), a convolutional neural network (CNN), rule-based regression (Cubist), Random Forest, and partial least squares regression (PLSR) using full spectra as well as features derived from principal component analysis (PCA) and partial least squares (PLS) latent variables. TabPFN consistently delivered the best overall performance across scales, including large spectral library tasks with tens of thousands of soil samples. Notably, TabPFN applied directly to full spectra already surpassed all classical baselines, showing that explicit dimensionality reduction is not strictly required for strong performance. Further improvements were achieved through PLS, which proved to be an effective dimensionality reduction strategy for all models. Combining PLS latent variables with TabPFN yielded the best predictions overall. Our findings provide evidence-based guidance for spectroscopic calibration model selection across operational scales, demonstrating that the long-standing advantages of PLSR and modern tabular foundation models complement each other in chemometrics.
