计算机科学

计算机科学 ›› 2026, Vol. 53 ›› Issue (10): 229-241.doi: 10.11896/jsjkx.250900067

• 数据库 & 大数据 & 数据科学 • 上一篇    下一篇

基于双自适应约束的对抗学习过采样方法

徐贞顺1,2, 郑顺国1, 苏梦瑶1, 赵泽宇1, 张耀欢1, 郭焱1   

  1. 1 北方民族大学计算机科学与技术学院 银川 750021
    2 北方民族大学图像图形智能处理国家民委重点实验室 银川 750021
  • 收稿日期:2025-09-11 修回日期:2025-12-08 发布日期:2026-10-13
  • 通讯作者: 徐贞顺(seojaein@qq.com)
  • 基金资助:
    北方民族大学高层次人才项目(自科)(2025BG240);北方民族大学研究生创新项目(YCX24359);宁夏重点研发计划(重点)项目(2023BDE02001);银川市校企联合创新项目(2022XQZD009)

Adversarial Learning with Dual Adaptive Constraints for Oversampling Method

XU Zhenshun1,2, ZHENG Shunguo1, SU Mengyao1, ZHAO Zeyu1, ZHANG Yaohuan1, GUO Yan1   

  1. 1 School of Computer Science and Technology,North Minzu University,Yinchuan 750021, China
    2 Key Laboratory of Image and Graphic Intelligent Processing,State Ethnic Affairs Commission,North Minzu University,Yinchuan 750021, China
  • Received:2025-09-11 Revised:2025-12-08 Online:2026-10-13
  • About author:XU Zhenshun,born in 1977,Ph.D,is a member of CCF(No.86794M).Her main research interests include data mining and knowledge discovery.
  • Supported by:
    High level Talent Project of North Minzu University(2025BG240),Graduate Innovation Program of North Minzu University(YCX24359),Key Research and Development Program of Ningxia(2023BDE02001) and School Enterprise Joint Innovation Project of Yinchuan City(2022XQZD009).

PDF (PC)

73

摘要： 分类算法在处理不平衡数据集时面临显著挑战,过采样方法通过合成新样本来平衡各类别样本数量,已成为主流解决方案之一。然而,现有方法在样本生成质量、高失衡率下模型的过度泛化以及特征学习稳定性方面存在显著局限。为此,提出基于双自适应约束的对抗学习过采样方法。该方法包括联合对抗学习框架和双自适应约束两部分。首先,联合对抗学习框架通过变分自编码器和判别器形成对抗训练,提高生成样本的质量和多样性;其次,采用差异自适应与标准自适应约束协同筛选,防止高失衡数据集中少数类样本的过度泛化和特征学习不稳定,确保生成样本逐步逼近真实分布并稳定合成。实验结果表明,该方法的F1-Score,G-mean和AUC等指标平均排名第一,其中在高失衡率数据集上的G-mean平均提升16.8%。所提方法有效提升了样本生成质量与特征表达能力,为不平衡数据分类中的关键问题提供了一种有效的解决方案。

关键词: 不平衡数据分类, 特征学习稳定性, 双自适应约束, 变分自编码器, 对抗学习

Abstract: Classification algorithms face significant challenges when dealing with imbalanced datasets.Oversampling methods,which synthesize new samples to balance the sample sizes across classes,have emerged as a mainstream solution.However,exis-ting approaches exhibit notable limitations in terms of the quality of generated samples,model over-generalization under high imbalance ratios,and instability in feature learning.To address these issues,this paper proposes an adversarial learning-based oversampling method with dual adaptive constraints.The method consists of two components:a joint adversarial learning framework and dual adaptive constraints.Firstly,the joint adversarial learning framework employs variational autoencoders and discriminators to form an adversarial training process,enhancing the quality and diversity of the generated samples.Secondly,differential adaptive and standard adaptive constraints are collaboratively applied to filter samples,preventing over-generalization and unstable feature learning in highly imbalanced datasets.This ensures that the generated samples progressively approximate the real distribution and are stably synthesized.Experimental results show that the proposed method ranks first on average in terms of F1-Score,G-mean,and AUC,achieving an average improvement of 16.8% in G-mean on highly imbalanced datasets.The proposed method effectively enhances sample generation quality and feature representation capability,providing an effective solution to key problems in imbalanced data classification.

Key words: Imbalanced data classification, Feature learning stability, Dual adaptive constraints, Variational autoencoder, Adversarial learning

中图分类号: 

  • TP391
[1] HE H,GARCIA E A.Learning from imbalanced data[J].IEEE Transactions on knowledge and data engineering,2009,21(9):1263-1284.
[2] GAN D,SHEN J,AN B,et al.Integrating TANBN with costsensitive classification algorithm for imbalanced data in medical diagnosis[J].Computers & Industrial Engineering,2020,140:106266.
[3] LI Z,HUANG M,LIU G,et al.A hybrid method with dynamic weighted entropy for handling the problem of class imbalance with overlap in credit card fraud detection[J].Expert Systems with Applications,2021,175:114750.
[4] KHLEEL N A A,NEHÉZ K.Software defect prediction using a bidirectional LSTM network combined with oversampling techniques[J].Cluster Computing,2024,27(3):3615-3638.
[5] RATHORE M S,HARSHA S P.Framework for imbalancedfault diagnosis of rolling bearing using autoencoding generative adversarial learning[J].Journal of the Brazilian Society of Mechanical Sciences and Engineering,2023,45(1):1806-3691.
[6] JIANG X,GE Z.Data augmentation classifier for imbalancedfault classification[J].IEEE Transactions on Automation Science and Engineering,2020,18(3):1206-1217.
[7] LI Y X,CHAI Y,HU Y Q,et al Overview of Unbalanced Data Classification Methods [J].Control and Decision,2019,34(4):673-688.
[8] GUO H,SUN T,LIU H,et al.Adaptive ensemble learning withcategory-aware attention and local contrastive loss[J].IEEE Transactions on Circuits and Systems for Video Technology,2024,35(2):1224-1236.
[9] WANG J Q,YUAN Y,ZHU Y T,et al.Ensemble Meta Net Based on Adaptive Reweight and Regularization [J].Application Research of Computers,2024,41(6):1749-1755.
[10] JING X Y,ZHANG X,ZHU X,et al.Multiset feature learning for highly imbalanced data classification[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2019,43(1):139-156.
[11] WANG X X,LI L X,LIN H.A Review of SMOTE Algorithm Research [J].Computer Science and Exploration,2024,18(5):1135-1159.
[12] CHAWLA N V,BOWYER K W,HALL L O,et al.SMOTE:synthetic minority over-sampling technique[J].Journal of Artificial Intelligence Research,2002,16:321-357.
[13] HAN H,WANG W Y,MAO B H.Borderline-SMOTE:A New Over-sampling Method in Imbalanced Data Sets Learning[C]//International Conference on Intelligent Computing.Berlin:Springer,2005:878-887.
[14] HE H,BAI Y,GARCIA E A,et al.ADASYN:Adaptive Synthetic Sampling Approach for Imbalanced Learning[C]//2008 IEEE International Joint Conference on Neural Networks(IEEE World Congress on Computational Intelligence).IEEE,2008:1322-1328.
[15] BARUA S,ISLAM M M,YAO X,et al.MWMOTE－majority weighted minority oversampling technique for imbalanced data set learning[J].IEEE Transactions on Knowledge and Data Engineering,2012,26(2):405-425.
[16] GOODFELLOW I,POUGET-A J,MIRZA M,et al.Generative adversarial nets[C]//Proceedings of the 28th International Conference on Neural Information Processing Systems.2014:2672-2680.
[17] KINGMA D P,WELLING M.Auto-encoding variational bayes[J].arXiv:1312.6114,2013.
[18] DOUZAS G,BACAO F.Effective data generation for imbal-anced learning using conditional generative adversarial networks[J].Expert Systems with Applications,2018,91:464-471.
[19] ENGELMANN J,LESSMANN S.Conditional wasserstein GAN-based oversampling of tabular data for imbalanced learning[J].Expert Systems with Applications,2021,174:114582.
[20] DING H,SUN Y,WANG Z,et al.RGAN-EL:A GAN and ensemble learning-based hybrid approach for imbalanced data classification[J].Information Processing & Management,2023,60(2):103235.
[21] GUAN S,ZHAO X,XUE Y,et al.AWGAN:An adaptiveweighting GAN approach for oversampling imbalanced datasets[J].Information Sciences,2024,663:120311.
[22] TAO X,ZHANG X,ZHENG Y,et al.A meanshift-guided oversampling with self-adaptive sizes for imbalanced data classification[J].Information Sciences,2024,672:120699.
[23] TIAN J,JIANG Y,ZHANG J,et al.A novel data augmentation approach to fault diagnosis with class-imbalance problem[J].Reliability Engineering & System Safety,2024,243:109832.
[24] XU L,SKOULARIDOU M,CUESTA-INFANTE A,et al.Modeling tabular data using conditional GAN.2019[J].arXiv:1907.00503,1907.
[25] DING Y,KANG W,FENG J,et al.Credit card fraud detection based on improved variational autoencoder generative adversarial network[J].IEEE Access,2023,11:83680-83691.
[26] ANSHELEVICH D,KATZ G.Synthetic tabular data generation using a VAE-GAN architecture[J].Knowledge-Based Systems,2025,326(C):113997.
[27] KANG F,FENG T,LIN J.VAE-GAN-Guided cross-class ge-neration:A class imbalance data augmentation method for network intrusion detection[J].Electronics,2025,14(11):2103.
[28] BENÍTEZ-PEÑA S,BLANQUERO R,CARRIZOSA E,et al.Cost-sensitive probabilistic predictions for support vector machines[J].European Journal of Operational Research,2024,314(1):268-279.
[29] LEEVY J L,JOHNSON J M,HANCOCK J,et al.Threshold optimization and random undersampling for imbalanced credit card data[J].Journal of Big Data,2023,10(1):58.
[30] LI S,PENG Y,SHEN Y,et al.Rolling bearing fault diagnosis under data imbalance and variable speed based on adaptive clustering weighted oversampling[J].Reliability Engineering & System Safety,2024,244:109938.
[31] LI S,SONG L,WU X,et al.Multi-class imbalance classification based on data distribution and adaptive weights[J].IEEE Transactions on Knowledge and Data Engineering,2024,36(10):5265-5279.
[32] WANG S,LIU W,YANG S,et al.An optimized adaBoost algorithm with atherosclerosis diagnostic applications:adaptive weight-adjustable boosting[J].The Journal of Supercomputing,2024,80(9):13187-13216.
[33] DONG Y,XIAO H,DONG Y.SA-CGAN:An oversamplingmethod based on single attribute guided conditional GAN for multi-class imbalanced learning[J].Neurocomputing,2022,472:326-337.
[34] ALCALA-FDEZ J,FERNANDEZ A,LUENGO J,et al.KEEL Data-Mining Software Tool:Data Set Repository,Integration of Algorithms and Experimental Analysis Framework[J].Journal of Multiple-Valued Logic and Soft Computing,2011,17(2/3):255-287.
[35] DOUZAS G,BACAO F,LAST F.Improving imbalanced learning through a heuristic oversampling method based on k-means and SMOTE[J].Information Sciences,2018,465:1-20.
[36] YAN Y,JIANG Y,ZHENG Z,et al.LDAS:local density-based adaptive sampling for imbalanced data classification[J].Expert Systems with Applications,2022,191:116213.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市两江新区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发