Accurate classification of dry bean varieties is important for automated agricultural inspection, quality assessment, and efficient crop management. This study proposes an Ant Colony Optimization-based Extreme Learning Machine (ACO-ELM) for the classification of dry bean varieties using morphological characteristics. The Dry Bean dataset initially contained 13,611 samples with 16 numerical features representing geometric and shape-related characteristics across seven bean classes. To minimize potential data leakage, 68 duplicate records were identified and removed before data partitioning, resulting in 13,543 unique samples. The dataset was divided using a stratified 80:20 train-test split, producing 10,834 training samples and 2,709 independent test samples. An internal validation subset was further created from the training data for ACO-based hyperparameter optimization. ACO was employed to optimize the number of hidden neurons, regularization parameter, and activation function of the ELM using 20 ants over 20 iterations. The optimization selected 900 hidden neurons, a regularization parameter (C = 50.0), and a sigmoid activation function, achieving a validation accuracy of 94.83%. The optimized ELM was subsequently retrained using the complete training set and evaluated on the independent test set. The proposed ACO-ELM achieved 92.47% accuracy, 92.47% weighted precision, 92.47% weighted recall, and 92.46% weighted F1-score. Furthermore, the model achieved a 99.05% weighted multiclass ROC-AUC, with macro precision, recall, and F1-score of 93.65%, 93.45%, and 93.54%, respectively. The results demonstrate that ACO-based hyperparameter optimization can provide an effective ELM-based framework for multiclass dry bean variety classification.
This study presented an Ant Colony Optimization-based Ex treme Learning Machine (ACO-ELM) for multiclass classifica tion of dry bean varieties using morphological characteristics. The Dry Bean dataset initially contained 13,611 samples and 16 numerical features representing geometric and shape-related properties of seven bean classes. To reduce the possibility of data leakage, 68 duplicate records were removed before data partitioning, resulting in 13,543 unique samples. A stratified 80:20 train-test strategy was adopted, while an internal val idation subset of the training data was used exclusively for ACO-based hyperparameter optimization. The proposed ACO framework optimized the number of hid den neurons, regularization parameter, and activation function of the ELM. The optimization process identified 900 hidden neurons, C = 50.0, and sigmoid activation as the best con figuration, achieving a validation accuracy of 94.83%. After optimization, the ELM was retrained using the complete train ing set and evaluated on an independent test set. The proposed ACO-ELM achieved 92.47% accuracy, 92.47% weighted preci sion, 92.47% weighted recall, and 92.46% weighted F1-score. In addition, the model obtained a 99.05% weighted multiclass ROC-AUC, demonstrating strong discriminative capability. The class-wise analysis showed strong performance for most bean varieties, with BOMBAY achieving 100% precision, re call, and F1-score. The problem was mainly encountered when classifying SIRA and DERMASON because of the similarity of their morphologies. Regardless of the problem, the high macro F1-score of 93.54% shows that the model still performed well on the seven classes. On the whole, the experimental results indicate that the ACO algorithm is efficient enough to tune hy perparameters of ELM and to offer a robust machine learning platform for classifying dry beans. It must be noted that the use of a nonlinear feature representation with the help of ELM com bined with population-based hyperparameter tuning with the help of ACO provides an efficient solution as compared to more sophisticated classifiers. The future research can explore cross validation-based tuning, feature selection algorithms, ensemble learning, among others.
References
[1] G. Slowinski, “Dry Beans Classification Using Machine Learning,” 29th International Workshop on Concurrency, Specification and Programming (CS&P’21), CEURWork shop Proceedings, 2021. 7 learning, among others. [2] R. Ardeshirifar, “Automated Classification of Dry Bean Varieties Using XGBoost and SVM Models,” arXiv:2408.01244v1 [cs.LG], Aug. 2, 2024. [3] A. Mehta, P. Sengupta, D. Garg, H. Singh, and Y. S. Diamand, “Benchmarking the Effectiveness of Classi fication Algorithms and SVM Kernels for Dry Beans,” arXiv:2307.07863, 2023. [4] M. S. Khan, T. D. Nath, M. S. Hossain, A. Mukherjee, H. B. Hasnath, T. M. Meem, and U. Khan, “Comparison of Multiclass Classification Techniques Using Dry Bean Dataset,” International Journal of Cognitive Computing in Engineering, 2023. [5] Q. Zaheer, A. U. Rehman, S. Javed, T. M. Ali, and A. Mir, “Leveraging Ensemble Learning for Dry Beans Clas sification,” International Conference on Engineering & Computing Technologies, May 2024. [6] N. K. Naik, P. K. Sethy, R. Amat, S. K. Behera, and P. Biswas, “Evaluation of Optimization Techniques with Support Vector Machine for Identification of Dry Beans,” Institute of Advanced Engineering and Science (IAES), vol. 32, no. 2, 2023. [7] V. Nasteski, “An Overview of the Supervised Machine Learning Methods,” Dec. 2017. [8] J. A. Ilemobayo, O. I. Durodola, T. O. Adewumi, O. Falana, et al., “Hyperparameter Tuning in Machine Learn ing: A Comprehensive Review,” Journal of Engineering Research and Reports, vol. 26, no. 6, pp. 388–395, June 2024. [9] P. Probst, A.-L. Boulesteix, and B. Bischl, “Tunability: Importance of Hyperparameters of Machine Learning Algorithms,” Journal of Machine Learning Research, vol. 20, pp. 1–32, 2019. [10] J. Bergstra and Y. Bengio, “Random Search for Hyper Parameter Optimization,” Journal of Machine Learning Research, vol. 13, pp. 281–305, 2012.
📋 How to Cite This Paper
MOHIT KHARBANDA, ARSHI HUSAIN, VIRENDRA P. VISHWAKARMA (2026). ACO-ELM: An Ant Colony Optimization-Driven Extreme Learning Machine for Dry Bean Variety Classification. International Journal of Engineering and Techniques, 12(4), 186–193. ISSN: 2395-1303. DOI: https://doi.org/10.5281/zenodo.22096805
