南京理工大学学报（自然科学版）2025，Vol.49Issue(2)：146-154,9.DOI:10.14177/j.cnki.32-1397n.2025.49.02.002
基于偏好学习的多准则分类问题建模及其ADMM研究
Research on modeling and ADMM for multi-criteria sorting problems based on preference learning
摘要
Abstract
Multi-criteria sorting problem is a research hotspot in the field of decision sciences and has wide applications in areas such as finance,education,and human resource management.Targeting this problem,a data-driven preference learning model is constructed based on additive value functions and linear approximation methods.The objective function characterizes the intra-cluster distance and inter-cluster distance.Besides,a sparse regularization term is introduced to enhance the model's generalization capability.Subsequently,an alternating direction method of multipliers(ADMM)is proposed to solve the model.Finally,numerical experiments and comparisons based on real-world datasets are conducted.The experimental results indicate that the proposed model effectively suppresses the overfitting phenomenon,achieving an out-of-sample classification accuracy of over 80%.Moreover,compared to traditional solvers,the proposed algorithm significantly improves computational efficiency.关键词
多准则分类问题/偏好学习/加性价值函数/正则化/交替方向乘子法Key words
multi-criteria sorting problem/preference learning/additive value function/regularization/alternating direction method of multipliers分类
数学引用本文复制引用
过燕晶,陈凯伦,吴中明..基于偏好学习的多准则分类问题建模及其ADMM研究[J].南京理工大学学报（自然科学版）,2025,49(2):146-154,9.基金项目
国家自然科学基金(12471291) （12471291）
江苏省自然科学基金(BK20241899) （BK20241899）
