The week's pick
Random Articles
Reseach Article
Application of Few Shot Learning using Prototypical Networks for Classification of Neglected Tropical Diseases in Low Resource Settings
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 187 - Number 14 |
| Year of Publication: 2025 |
| Authors: Akinrotimi Akinyemi Omololu, Atoyebi Jelili Olaniyi, Owolabi Olugbenga Olayinka |
10.5120/ijca2025925176
|
Akinrotimi Akinyemi Omololu, Atoyebi Jelili Olaniyi, Owolabi Olugbenga Olayinka . Application of Few Shot Learning using Prototypical Networks for Classification of Neglected Tropical Diseases in Low Resource Settings. International Journal of Computer Applications. 187, 14 ( Jun 2025), 66-73. DOI=10.5120/ijca2025925176
Abstract
Neglected tropical diseases (NTDs) remain a substantial public health issue in low-resource settings such as Ayedaade Local Government Area, Osun State, Nigeria. These settings often lack properly labeled medical data, making traditional machine learning approaches inapplicable. In this paper, the use of Few-Shot Learning (FSL), specifically Prototypical Networks, for the identification of three major NTDs, viz., Lymphatic Filariasis, Onchocerciasis, and Schistosomiasis, is described. The performance of the FSL model was evaluated against baseline classifiers including Naive Bayes, K-Nearest Neighbors, and Random Forest. Prototypical Networks demonstrated the highest accuracy at 87.7%, followed by Random Forest at 81.7%, K-Nearest Neighbors at 78.2%, and Naive Bayes at 72.5%. On individual disease classification, 90.3% accuracy, 0.88 F1-Score, and 0.94 AUC-ROC were achieved in the case of Schistosomiasis. Comprehensive visualization bar graph, line chart and a confusion matrix, have been used to display model performance on all values. The performance of the FSL model was evaluated against baseline classifiers including Naive Bayes, K-Nearest Neighbors, and Random Forest. Prototypical Networks demonstrated the highest accuracy at 87.7%, followed by Random Forest at 81.7%, K-Nearest Neighbors at 78.2%, and Naive Bayes at 72.5%.
References
- O. Vinyals, C. Blundell, T. Lillicrap, D. Wierstra, and K. Kavukcuoglu, "Matching Networks for One-Shot Learning," Proc. NeurIPS 2016, vol. 29, pp. 3637-3645, 2016. [Online]. Available: https://arxiv.org/abs/1606.04080.
- J. Snell, K. Swersky, and R. Zemel, "Prototypical Networks for Few-Shot Learning," Proc. NeurIPS 2017, vol. 29, pp. 4077-4087, 2017. [Online]. Available: https://arxiv.org/abs/1703.05175.
- X. Li, H. Wang, T. Zhang, and Z. Liu, "Few-shot learning in medical imaging: A survey," Med. Image Anal., vol. 65, p. 101767, 2020. [Online]. Available: https://doi.org/10.1016/j.media.2020.101767.
- Y. Wang, L. Zhang, and W. Liu, "Few-shot learning for rare disease prediction: A deep learning approach," IEEE Trans. Biomed. Eng., vol. 68, no. 4, pp. 1102-1110, 2021. [Online]. Available: https://doi.org/10.1109/TBME.2020.2990420.
- Li, X., Zhang, Y., & Liu, J., “Few-shot learning techniques for medical image analysis,” Medical Image Analysis, vol. 34, no. 2, pp. 112-120, 2020.
- Snell, J., Swersky, K., & Zemel, R., “Prototypical networks for few-shot learning,” in Proceedings of Neural Information Processing Systems (NeurIPS), 2017, pp. 4077-4087.
- Wang, L., Zhang, Z., & Li, F., “Deep learning for rare disease prediction from tabular clinical data,” IEEE Transactions on Medical Imaging, vol. 42, no. 8, pp. 987-994, 2021.
- Zhang, Y., Wang, L., & Liu, Y., “Deep learning methods in rare disease prediction,” Medical Data Science Journal, vol. 28, no. 3, pp. 105-112, 2020.
- Jiang, H., Chen, Y., & Xu, X., “Few-shot learning for clinical datasets,” Journal of Healthcare Informatics Research, vol. 15, no. 1, pp. 45-52, 2021.
- F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011.
- A. Géron, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems, 2nd ed. Sebastopol, CA: O’Reilly Media, 2019.
Index Terms
Keywords