计算机科学 ›› 2023, Vol. 50 ›› Issue (10): 88-95.doi: 10.11896/jsjkx.230600048
何玉林1,2, 朱鹏辉2, 黄哲学1,2, Fournier-Viger PHILIPPE2
HE Yulin1,2, ZHU Penghui2, HUANG Zhexue1,2, Fournier-Viger PHILIPPE2
摘要: 半监督集成是将半监督学习与集成学习相结合的一种学习范式,它一方面通过无标记样本来提高集成学习的多样性,同时解决集成学习样本量不足的问题,另一方面集成多个分类器能够进一步提升半监督学习模型的性能。现有的研究从理论和实践两个角度证明了半监督学习与集成学习之间的互益性。针对当前半监督集成学习算法对无标记样本信息利用不完全的缺陷,文中提出了一种新的基于分类不确定性最小化的半监督集成学习(Classification Uncertainty Minimization-Based Semi-Supervised Ensemble Learning,CUM-SSEL)算法,它引入信息熵作为对无标记样本进行打标的置信度评判标准,通过最小化无标记样本打标过程中的不确定性迭代地训练分类器,实现对无标记样本的高效利用,以增强分类器的泛化性能。在标准的实验数据集上对CUM-SSEL算法的可行性、合理性和有效性进行了验证,实验表明:随着基分类器的增加,CUM-SSEL算法的训练呈现收敛的趋势,同时它能够获得优于Self-Training,Co-Training,Tri-Training,Semi-Boost,Vote-Training,Semi-Bagging以及CST-Voting算法的分类精度。
中图分类号:
[1]MERZ C,CLAIR D S,BOND W.Semi-supervised adaptive resonance theory[C]//Proceedings of IJCNN International Joint Conference on Neural Networks.IEEE,1992,3:851-856. [2]HADY M,SCHWENKER F.Semi-Supervised Learning [J].Journal of the Royal Statistical Society,2006,172(2):530. [3]VAN ENGELEN J,HOOS H H.A survey on semi-supervisedlearning [J].MachineLearning,2020,109(2):373-440. [4]BUHLMANN P,YU B.Analyzing bagging [J].Annals of Sta-tistics,2002,30(4):927-961. [5]SCHAPIRE R E.The boosting approach to machine learning:An overview [J].Lecture Notes in Statistics:Nonlinear Estimation and Classification,2003,171:149-171. [6]SAGI O,ROKACH L.Ensemble learning:A survey[J].Wiley Interdisciplinary Reviews:Data Mining and Knowledge Disco-very,2018,8(4):e1249. [7]BENNETT K P,DEMIRIZ A,MACLIN R.Exploiting unla-beled data in ensemble methods[C]//Proceedings of the eighth ACM SIGKDD International Conference on Knowledge Disco-very and Data Mining.2002:289-296. [8]ZHOU Z H.When semi-supervised learning meets ensemblelearning[J].Frontiers of Electrical and Electronic Engineering in China,2011,6:6-16. [9]DONG X,YU Z,CAO W,et al.A survey on ensemble learning[J].Frontiers of Computer Science,2020,14:241-258. [10]BREIMAN L.Bagging predictors [J].Machine Learning,1996,24:123-140. [11]SEEDAT N,KANAN C.Towards calibrated and scalable uncertainty representations for neural networks [J].arXiv:1911.00104,2019. [12]MALLAPRAGADA P K,JIN R,JAIN A K,et al.Semiboost:Boosting for semi-supervised learning [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,31(11):2000-2014. [13]LUO Y,ZHU J,LI M,et al.Smooth neighbors on teachergraphs for semi-supervised learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2018:8896-8905. [14]HOU J,MAO Y,SUN J S.A Semi-supervised Boosting Algorithm for Maximizing Sample Separability [J].Journal of Nanjing University of Technology,2014,38(5):675-681. [15]YANG J,ZHANG D,YANG J Y,et al.Globally maximizing,locally minimizing:unsupervised discriminant projection with applications to face and palm biometrics [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2007,29(4):650-664. [16]TANHA J.MSSBoost:A new multiclass boosting to semi-supervised learning [J].Neurocomputing,2018,314:251-266. [17]CHEN S,SU S,LI S Z,et al.Cooperative training target tra-cking algorithm based on online semi-supervised boosting [J].Journal of Electronics and Information,2014,36(4):888-895. [18]ZHANG M L,ZHOU Z H.Exploiting unlabeled data to enhance ensemble diversity [J].Data Mining and Knowledge Discovery,2013,26:98-129. [19]LI Y,SU L,CHEN J,et al.Semi-supervised learning for question classification in CQA [J].Natural Computing,2017,16:567-577. [20]LIVIERIS I E,KANAVOS A,TAMPAKAS V,et al.An en-semble SSL algorithm for efficient chest X-ray image classification [J].Journal of Imaging,2018,4(7):95. [21]YAROWSKY D.Unsupervised word sense disambiguation rivaling supervised methods[C]//33rd Annual Meeting of The Association for Computational Linguistics.1995:189-196. [22]BLUM A,MITCHELL T.Combining labeled and unlabeled data with co-training[C]//Proceedings of The Eleventh Annual Conference on Computational Learning Theory.1998:92-100. [23]ZHOU Z H,LI M.Tri-training:Exploiting unlabeled data using three classifiers [J].IEEE Transactions on Knowledge and Data Engineering,2005,17(11):1529-1541. [24]KENDALL A,GAL Y.Whatuncertainties do we need in Baye-sian deep learning for computer vision?[C]//Proceedings of the 31st Conference on Neural Information Processing Systems.2017:5580-5590. [25]GE J,MA T.Semi-supervised learning based on ensemble algorithm[C]//Proceedings of the 29th China Database Academic Conference.2012:208-213. |
|