Interpretable Machine Learning for Crop Classification Using Decision Boundary Visualization of Multi Classifier Models
Journal of Smart and Sustainable Farming
Received On : 28 December 2025
Revised On : 30 January 2026
Accepted On : 18 February 2026
Published On : 26 February 2026
Volume 02, 2026
Pages : 022-032
Abstract
Proper classification of crops is one of the core aspects of precision agriculture since it enhances yield prediction, management of resources, and decision-making. Nonetheless, the current machine learning methods are usually confronted with the problem of achieving good classification accuracy versus interpretability, especially in the case of complex and nonlinear agricultural data. SVM and other traditional models that include Support Vector machine and Random Forest offer good competition but do not give any insight as to how decisions are made, thus restricting their use in practice. In order to overcome these weaknesses, this paper suggests a new interpretable model which it has called ADB-Adaptive Decision Boundary Agro Net (AgroNet) to classify crops. The model proposed is a combination of the polynomial expansion of features and a neural network-based classifier that will effectively learn the interactions of features of higher order and increase the separability of the classes. Moreover, a decision boundary visualization based on Principal Component Analysis is included, allowing the human user to have a feel of how the model works, which does not affect the training. The combination of feature engineering and adaptive learning enables the model to produce more distinguishable and smooth decision regions than the traditional methods. ADB-AgroNet performance is compared to models of the state-of-the-art, such as SVM, Random Forest, and K-Nearest Neighbors, with the help of the extensive set of evaluation criteria. It is shown in the experiments that the proposed model has a better classification performance with an accuracy of 94.82, and higher precision, recall, and F1-score. Moreover, the model has a greater robustness and a lower rate of misclassification among all classes. The results indicate that ADB-AgroNet is a practical and reliable tool of solving an issue of accuracy and interpretability that can be applied into the real-life agricultural context.
Keywords
ADB (Adaptive Decision Boundary), AgroNet, (SVM) Support Vector Machine, Multi-Layer Perceptrons (MLPs).
- A. Mohagheghi and M. Moallem, “IoT-Based Plant Identification Using Multi-Level Classification,” IEEE Access, vol. 12, pp. 146366–146375, 2024, doi: 10.1109/access.2024.3474613.
- T. Rusňák et al., “Crop Mapping without Labels: Investigating Temporal and Spatial Transferability of Crop Classification Models Using a 5-Year Sentinel-2 Series and Machine Learning,” Remote Sensing, vol. 15, no. 13, p. 3414, Jul. 2023, doi: 10.3390/rs15133414.
- S. Albahli and M. Masood, “Efficient attention-based CNN network (EANet) for multi-class maize crop disease classification,” Frontiers in Plant Science, vol. 13, Oct. 2022, doi: 10.3389/fpls.2022.1003152.
- D. Goel, Anshul, G. Chauhan, A. Singh, and M. Chaudhry, “Enchancing Crop Productivity through Early Plant Disease Identification,” 2024 International Conference on Computational Intelligence and Computing Applications (ICCICA), pp. 226–230, May 2024, doi: 10.1109/iccica60014.2024.10584982.
- A. J. Rivera, M. D. Pérez-Godoy, D. Elizondo, L. Deka, and M. J. del Jesus, “Analysis of clustering methods for crop type mapping using satellite imagery,” Neurocomputing, vol. 492, pp. 91–106, Jul. 2022, doi: 10.1016/j.neucom.2022.04.002.
- A. Elkassar, “Deep Learning based Grapevine Leaf Classification using Augmented Images and Multi-Classifier Fusion for Improved Accuracy and Precision,” 2024 14th International Conference on Electrical Engineering (ICEENG), pp. 190–192, May 2024, doi: 10.1109/iceeng58856.2024.10566412.
- Y. Lv, W. Feng, S. Wang, S. Wang, L. Guo, and G. Dauphin, “An Ensemble-Based Framework for Sophisticated Crop Classification Exploiting Google Earth Engine,” Remote Sensing, vol. 16, no. 5, p. 917, Mar. 2024, doi: 10.3390/rs16050917.
- R. Simeón et al., “Machine Learning Applied to Crop Mapping in Rice Varieties Using Spectral Images,” Agriculture, vol. 15, no. 17, p. 1832, Aug. 2025, doi: 10.3390/agriculture15171832.
- A. S. Keceli, A. Kaya, C. Catal, and B. Tekinerdogan, “Deep learning-based multi-task prediction system for plant disease and species detection,” Ecological Informatics, vol. 69, p. 101679, Jul. 2022, doi: 10.1016/j.ecoinf.2022.101679.
- D. Wadhwa and K. Malik, “Transfer Learning for Generalized Multi-Species Crop Disease Classification: A Comparative Analysis of Fine-Tuned CNN Models,” Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics, pp. 55–68, Oct. 2025, doi: 10.1007/978-981-96-6843-4_5.
- W. Zhang et al., “Evaluating intra-season multi-temporal data organization strategies for land cover classification in irrigation districts using multimodal remote sensing data,” International Journal of Remote Sensing, pp. 1–40, Mar. 2026, doi: 10.1080/01431161.2026.2637849.
- S. Fathololoumi, H. B. Vasava, D. Saurette, P. Daggupati, and A. Biswas, “Enhanced ephemeral gully mapping through multi-classifier integration and spectral feature analysis,” CATENA, vol. 248, p. 108590, Jan. 2025, doi: 10.1016/j.catena.2024.108590.
- A. A. Kindalkar, “A Research Review and Perspective Toward Plant Leaf Disease Detection Using Image Processing Techniques,” International Journal of Image and Graphics, vol. 26, no. 04, Oct. 2024, doi: 10.1142/s0219467826500324.
- K. Chureesampant and J. Susaki, “Land cover classification using multi-temporal SAR data and optical data fusion with adaptive training sample selection,” 2012 IEEE International Geoscience and Remote Sensing Symposium, pp. 6177–6180, Jul. 2012, doi: 10.1109/igarss.2012.6352667.
- M. Anokye et al., “Optimizing multi-classifier fusion for seabed sediment classification using machine learning,” International Journal of Digital Earth, vol. 17, no. 1, Dec. 2023, doi: 10.1080/17538947.2023.2295988.
CRediT Author Statement
The author reviewed the results and approved the final version of the manuscript.
Acknowledgements
We would like to thank Reviewers for taking the time and effort necessary to review the manuscript. We sincerely appreciate all valuable comments and suggestions, which helped us to improve the quality of the manuscript.
Funding
No funding was received to assist with the preparation of this manuscript.
Ethics Declarations
Conflict of interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.
Availability of Data and Materials
The dataset used in this study is publicly available and can be accessed from the Kaggle Crop Recommendation Dataset repository. This dataset contains essential agricultural parameters such as nitrogen (N), phosphorus (P), potassium (K), temperature, humidity, pH, and rainfall, which are widely used for crop classification tasks. The dataset is available at: https://www.kaggle.com/datasets/atharvaingle/crop-recommendation-dataset All preprocessing steps, feature engineering (including polynomial feature expansion), and train-test splits were performed by the authors. The processed data and implementation code supporting the findings of this study can be made available from the corresponding author upon reasonable request.
Author Information
Contributions
All authors have equal contribution in the paper and all authors have read and agreed to the published version of the manuscript.
Corresponding Author
Rights and permissions
This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0).
https://creativecommons.org/licenses/by/4.0/.
This license permits unrestricted use, sharing, distribution, reproduction, and adaptation in any medium or format, including for commercial purposes, provided that appropriate credit is given to the original author(s) and the source, a link to the license is provided, and any changes made are indicated.
Copyright
© 2026 Takeru Kobayashi. The author(s) retain copyright of the work. The author(s) grant the Journal of Smart and Sustainable Farming (JSSF) and its publisher, Ansis Publications, the right of first publication and the right to identify itself as the original publisher of the article.
Cite this Article
Takeru Kobayashi, “Interpretable Machine Learning for Crop Classification Using Decision Boundary Visualization of Multi Classifier Models”, Journal of Smart and Sustainable Farming, pp. 022-032, 2026, doi: 10.64026/JSSF/2026003.