EUVIMEDEUVIMED entdecken
Crossref

An Explainable Stacked Ensemble Learning Framework for Crop Recommendation using Soil and Climatic Parameters

Fen Huang, Benjamin Hutchinson

Journal of Smart and Sustainable Farming · 02.06.2026 · 103

DOI: 10.64026/jssf/2026010 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Smart and Sustainable Farming, online erschienen am 02.06.2026. Verfasst von Fen Huang und 1 weiteren Autor:innen. Die Arbeit nennt 18 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Choosing the right crop for a piece of land relies on soil nutrients, soil acidity and local climate and if the wrong crop is selected, the yield will not be as good and inputs will be wasted. In this paper, an explainable stacked ensemble learning (EELE) model is presented which recommends one of the 22 crops based on seven soil and climatic parameters, namely, Nitrogen, Phosphorus, Potassium, Temperature, Relative Humidity, Soil pH and Rainfall. The base layer consists of random forest, histogram-based gradient boosting and a support vector machine with an RBF kernel, with the out-of-fold class probabilities being combined by a meta-learner, a multinomial logistical-regression. Permutation importance is introduced to indicate which parameters are important for a recommended classification. The system is tested on a test dataset consisting of 2200 samples generated by simulation from the ranges of agronomic requirements and is compared against eight classifiers serving as baseline classifiers. The held-out test set shows that the proposed model achieves 94.55% accuracy and 94.56% macro average F1 score, surpassing all baseline models, while a macro averaged AUC of 0.9994. It ranks second to a tuned SVM (94.59%) under 5×2 repeated cross validation and an ablation study reveals that the Gradient Boosting member makes the least contribution. The most important inputs are humidity, nitrogen and phosphorus. Learning Curves, noise robustness and computational cost are also reported and it discusses the limitations of what can be concluded from simulated data.

Zitieren

APA

Huang, F., & Hutchinson, B. (2026). An Explainable Stacked Ensemble Learning Framework for Crop Recommendation using Soil and Climatic Parameters. Journal of Smart and Sustainable Farming, 103. https://doi.org/10.64026/jssf/2026010

Vancouver

Huang F, Hutchinson B. An Explainable Stacked Ensemble Learning Framework for Crop Recommendation using Soil and Climatic Parameters. Journal of Smart and Sustainable Farming. 2026:103. doi:10.64026/jssf/2026010

BibTeX
@article{Huang2026,
  author = {Huang, Fen and Hutchinson, Benjamin},
  title = {An Explainable Stacked Ensemble Learning Framework for Crop Recommendation using Soil and Climatic Parameters},
  journal = {Journal of Smart and Sustainable Farming},
  year = {2026},
  pages = {103},
  doi = {10.64026/jssf/2026010}
}
Kostenlos starten →