有序分類廣泛應用於醫學診斷、年齡預測及信用評分等實務領域。然而，此類任務常面臨嚴重的類別不平衡問題，導致模型難以準確識別少數類別。儘管如此，多數現有有序分類方法未針對此挑戰進行專門設計，限制了其在實務中的效能與應用潛力。
為解決上述問題，我們提出 RankCL，一個專為不平衡有序分類任務設計的可插拔模組。RankCL 提出排序感知的對比學習，一種監督式對比學習機制，藉由標籤間的內在順序關係，引導樣本在嵌入空間中學習具備順序結構的表徵，提升有序分類能力。此外，其對比集合與相近關係學習策略，有助於強化少數類樣本的語意資訊，並且避免少數類樣本在訓練過程中被忽略，從而應對類別不平衡帶來的挑戰。
我們在多個真實世界有序分類資料集上進行了實證研究。實驗結果表明，RankCL 能夠整合至多種深度有序分類方法，提升多數方法的順序性與分類效能。特別是結合 OBDECOC 預測頭時，RankCL 在所有比較方法中取得最佳平均排名。

Ordinal classification is widely applied in practical domains such as medical diagnosis, age estimation, and credit scoring. However, these tasks often face severe class imbalance issues, making it difficult for models to accurately identify minority classes. Despite this, most existing ordinal classification methods are not specifically designed to address this challenge, limiting their effectiveness and practical applicability.
To overcome these issues, we propose RankCL, a plug-and-play module specially designed for imbalanced ordinal classification tasks. RankCL leverages ranking-aware contrastive learning, a supervised contrastive mechanism that leverages the inherent ordinal relationships among labels to guide samples in the embedding space to learn representations with preserved order structure, enhancing ordinal classification performance. Additionally, its contrastive set construction and similarity relation learning strategies help strengthen the semantic information of minority class samples and prevent them from being overlooked during training, thereby addressing the challenges caused by class imbalance.
We conduct empirical studies on multiple real-world ordinal classification datasets. Experimental results show that RankCL can be integrated with various deep ordinal classification methods, improving the ordinal consistency and classification performance of most approaches. Notably, when combined with the OBDECOC prediction head, RankCL achieves the best average ranking among all compared methods.

[1] A. Agresti, “Categorical data analysis. john wiley & sons,” Inc., Publication, vol. 15, p. 24, 2002.
[2] T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama, “Optuna: A next-generation hyperparameter optimization framework,” in Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, 2019, pp. 2623–2631.
[3] S. Baccianella, A. Esuli, and F. Sebastiani, “Evaluation measures for ordinal regression,”in 2009 Ninth international conference on intelligent systems design and applications. IEEE, 2009, pp. 283–287.
[4] J. Barbero-G´omez, P. A. Guti´errez, and C. Herv´as-Mart´ınez, “Error-correcting output codes in the framework of deep ordinal classification,” Neural Processing Letters, vol. 55, no. 5, pp. 5299–5330, 2023.
[5] F. B´erchez-Moreno, R. Ayll´on-Gavil´an, V. M. Vargas, D. Guijo-Rubio, C. Herv´as- Mart´ınez, J. C. Fern´andez, and P. A. Guti´errez, “dlordinal: A python package for deep ordinal classification,” Neurocomputing, p. 129305, 2025.
[6] F. B´erchez-Moreno, J. C. Fern´andez, C. Herv´as-Mart´ınez, and P. A. Guti´errez, “Fusion of standard and ordinal dropout techniques to regularise deep models,” Information Fusion, vol. 106, p. 102299, 2024.
[7] L. Breiman, “Random forests,” Machine learning, vol. 45, no. 1, pp. 5–32, 2001.
[8] J. C. F. Caballero, F. J. Mart´ınez, C. Herv´as, and P. A. Guti´errez,“Sensitivity versus accuracy in multiclass problems using memetic pareto evolutionary neural networks,”IEEE Transactions on Neural Networks, vol. 21, no. 5, pp. 750–770, 2010.
[9] W. Cao, V. Mirjalili, and S. Raschka, “Rank consistent ordinal regression for neural networks with application to age estimation,” Pattern Recognition Letters, vol. 140, pp.325–331, 2020.
[10] F. Castagnos, M. Mihelich, and C. Dognin, “A simple log-based loss function for ordinal text classification,” in Proceedings of the 29th International Conference on Computational Linguistics, 2022, pp. 4604–4609.
[11] T. Chen and C. Guestrin, “Xgboost: A scalable tree boosting system,” in Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, 2016, pp. 785–794.
[12] W. Chu and S. S. Keerthi, “Support vector ordinal regression,” Neural computation, vol. 19, no. 3, pp. 792–815, 2007.
[13] D. R. Cox, “The regression analysis of binary sequences,” Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 20, no. 2, pp. 215–232, 1958.
[14] M. Cruz-Ram´ırez, C. Herv´as-Mart´ınez, J. S´anchez-Monedero, and P. A. Guti´errez,“Metrics to guide a multi-objective evolutionary algorithm for ordinal classification,”Neurocomputing, vol. 135, pp. 21–31, 2014.
[15] J. de La Torre, D. Puig, and A. Valls, “Weighted kappa loss function for multi-class classification of ordinal data in deep learning,” Pattern Recognition Letters, vol. 105, pp. 144–154, 2018.
[16] R. Diaz and A. Marathe, “Soft labels for ordinal regression,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 4738–4747.
[17] I. Domingues, J. P. Amorim, P. H. Abreu, H. Duarte, and J. Santos, “Evaluation of oversampling data balancing techniques in the context of ordinal classification,” in 2018 International Joint Conference on Neural Networks (IJCNN). IEEE, 2018, pp. 1–8.
[18] M. Dorado-Moreno, M. P´erez-Ortiz, P. A. Guti´errez, R. Ciria, J. Brice˜no, and C. Herv´as-Mart´ınez, “Dynamically weighted evolutionary ordinal neural network for solving an imbalanced liver transplantation problem,” Artificial Intelligence in Medicine, vol. 77, pp. 1–11, 2017.
[19] F. Fern´andez-Navarro, “A generalized logistic link function for cumulative link models in ordinal regression,” Neural Processing Letters, vol. 46, no. 1, pp. 251–269, 2017.
[20] E. Frank and M. Hall, “A simple approach to ordinal classification,” in Machine Learning: ECML 2001: 12th European Conference on Machine Learning Freiburg, Germany, September 5–7, 2001 Proceedings 12. Springer, 2001, pp. 145–156.
[21] P. A. Guti´errez, M. Perez-Ortiz, J. Sanchez-Monedero, F. Fernandez-Navarro, and C. Hervas-Martinez, “Ordinal regression methods: survey and experimental study,”IEEE Transactions on Knowledge and Data Engineering, vol. 28, no. 1, pp. 127–146, 2015.
[22] G. Ke, Q. Meng, T. Finley, T.Wang,W. Chen,W. Ma, Q. Ye, and T.-Y. Liu, “Lightgbm: A highly efficient gradient boosting decision tree,” Advances in neural information processing systems, vol. 30, 2017.
[23] S. B. Kotsiantis and P. E. Pintelas, “A cost sensitive technique for ordinal classification problems,” in Hellenic Conference on Artificial Intelligence. Springer, 2004, pp. 220–229.
[24] S. Kramer, G.Widmer, B. Pfahringer, and M. De Groeve, “Prediction of ordinal classes using regression trees,” Fundamenta Informaticae, vol. 47, no. 1-2, pp. 1–13, 2001.
[25] M. L´azaro and A. R. Figueiras-Vidal, “Neural network for ordinal classification of imbalanced data by minimizing a bayesian cost,” Pattern Recognition, vol. 137, p. 109303, 2023.
[26] L. Li and H.-T. Lin, “Ordinal regression by extended binary classification,” Advances in neural information processing systems, vol. 19, 2006.
[27] X. Liu, F. Fan, L. Kong, Z. Diao, W. Xie, J. Lu, and J. You, “Unimodal regularized neuron stick-breaking for ordinal classification,” Neurocomputing, vol. 388, pp. 34–44, 2020.
[28] Y. Liu, A. W.-K. Kong, and C. K. Goh, “Deep ordinal regression based on data relationship for small datasets.” in IJCAI, 2017, pp. 2372–2378.
[29] Y. Liu, A. W. K. Kong, and C. K. Goh, “A constrained deep neural network for ordinal regression,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 831–839.
[30] F. Marques, H. Duarte, J. Santos, I. Domingues, J. P. Amorim, and P. H. Abreu, “An iterative oversampling approach for ordinal classification,” in Proceedings of the 34th ACM/SIGAPP symposium on applied computing, 2019, pp. 771–774.
[31] Z. Niu, M. Zhou, L. Wang, X. Gao, and G. Hua, “Ordinal regression with multiple output cnn for age estimation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 4920–4928.
[32] M. Perez-Ortiz, P. A. Gutierrez, C. Hervas-Martinez, and X. Yao, “Graph-based approaches for over-sampling in the context of ordinal regression,” IEEE Transactions on Knowledge and Data Engineering, vol. 27, no. 5, pp. 1233–1245, 2014.
[33] M. P´erez-Ortiz, A. S´aez, J. S´anchez-Monedero, P. A. Guti´errez, and C. Herv´as-Mart´ınez, “Tackling the ordinal and imbalance nature of a melanoma image classification problem,” in 2016 international joint conference on neural networks (IJCNN). IEEE, 2016, pp. 2156–2163.
[34] D. Pitawela, G. Carneiro, and H.-T. Chen, “Cloc: Contrastive learning for ordinal classification with multi-margin n-pair loss,” in Proceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 15 538–15 548.
[35] J. R. Quinlan, “Induction of decision trees,” Machine learning, vol. 1, no. 1, pp. 81–106, 1986.
[36] D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning representations by backpropagating errors,” nature, vol. 323, no. 6088, pp. 533–536, 1986.
[37] J. S´anchez-Monedero, P. A. Guti´errez, and M. P´erez-Ortiz, “Orca: A matlab/octave toolbox for ordinal regression,” Journal of Machine Learning Research, vol. 20, no. 125, pp. 1–5, 2019. [Online]. Available: http://jmlr.org/papers/v20/18-349.html
[38] F. Sebastiani, “Machine learning in automated text categorization,” ACM computing surveys (CSUR), vol. 34, no. 1, pp. 1–47, 2002.
[39] X. Shi, W. Cao, and S. Raschka, “Deep neural networks for rank-consistent ordinal regression based on conditional probabilities,” Pattern Analysis and Applications, vol. 26, no. 3, pp. 941–955, 2023.
[40] V. M. Vargas, A. M. G´omez-Orellana, P. A. Guti´errez, C. Herv´as-Mart´ınez, and D. Guijo-Rubio, “Ebano: A novel ensemble based on unimodal ordinal classifiers for the prediction of significant wave height,” Knowledge-Based Systems, vol. 300, p. 112223, 2024.
[41] V. M. Vargas, P. A. Guti´errez, and C. Hervas-Martinez, “Cumulative link models for deep ordinal classification,” Neurocomputing, vol. 401, pp. 48–58, 2020.
[42] T. Zhu, Y. Lin, Y. Liu,W. Zhang, and J. Zhang, “Minority oversampling for imbalanced ordinal regression,” Knowledge-Based Systems, vol. 166, pp. 140–155, 2019.