Group invariance is a crucial type of prior knowledge often employed to enhance learning performance. As a binary classification support vector machine algorithm, twin support vector machines (TWSVM) can improve performance by exploring group invariance. Thus, in this paper, we propose to incorporate group invariance into the framework of TWSVM and thereby define the problem of group invariance-based twin support vector machines (GI⁃TWSVM) to improve the performance. First, an optimization problem is formulated for GI⁃TWSVM. Using the Twin Bounded Support Vector Machine (TBSVM) as an example, we develop two TBSVM algorithms that incorporate group invariance, demonstrating that the optimization problem is solvable and practically significant. Then, we systematically investigate the consistency of GI⁃TWSVM to build a solid theoretical basis for the related algorithms. Finally, experimental results using TBSVM as an example indicate that group invariance can significantly enhance the performance of twin support vector machine algorithms.
LAUERF, BLOCHG. Incorporating prior knowledge in support vector machines for classification: A review[J]. Neurocomputing, 2008, 71(7/8/9): 1578-1594. DOI: 10.1016/j.neucom.2007.04.010 .
[2]
ZHANGW, YUL, YOSHIDAT, et al. Feature weighted confidence to incorporate prior knowledge into support vector machines for classification[J]. Knowledge and Information Systems, 2019, 58(2): 371-397. DOI: 10.1007/s10115-018-1165-2 .
[3]
LIUY T, HUANGL, LIJ, et al. Multi-task learning based on geometric invariance discriminative features[J]. Applied Intelligence, 2023, 53(3): 3505-3518. DOI: 10.1007/s10489-022-03617-x .
[4]
NANDAV, MAJUMDARA, KOLLINGC, et al. Do invariances in deep neural networks align with human perception?[J]. Proceedings of the 37th AAAI Conference on Artificial Intelligence, 2023, 37(8): 9277-9285. DOI: 10.1609/aaai.v37i8.26112 .
[5]
SIMARDP, LECUNY, DENKERJ S. Efficient pattern recognition using a new transformation distance[EB/OL]. [2022-12-25]. DOI: 10.1007/3-540-49430-8_13 .
[6]
HOUNIEI, CHAMONL F O, RIBEIROA. Automatic data augmentation via invariance-constrained learning[C]// Proceedings of the 40th International Conference on Machine Learning. New York: PMLR, 2023: 13410-13433. DOI: 10.5555/3618408.3618953 .
[7]
XUW X, HUANGD J, ZHOUS G. Statistical learning with group invariance: Problem, method and consistency[J]. International Journal of Machine Learning and Cybernetics, 2019, 10(6): 1503-1511. DOI: 10.1007/s13042-018-0829-2 .
[8]
JAYADEVA, KHEMCHANDANIR, CHANDRAS. Twin support vector machines for pattern classification[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 29(5): 905-910. DOI: 10.1109/TPAMI.2007.1068 .
[9]
MANGASARIANO L, WILDE W. Multisurface proximal support vector machine classification via generalized eigenvalues[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(1): 69-74. DOI: 10.1109/TPAMI.2006.17 .
[10]
SHAOY H, ZHANGC H, WANGX B, et al. Improvements on twin support vector machines[J]. IEEE Transactions on Neural Networks, 2011, 22(6): 962-968. DOI: 10.1109/TNN.2011.2130540 .
[11]
SHAOY H, CHENW J, WANGZ, et al. Weighted linear loss twin support vector machine for large-scale classification[J]. Knowledge–Based Systems, 2015, 73(1): 276-288. DOI: 10.1016/j.knosys.2014.10.011 .
[12]
TANVEERM, SHUBHAMK. Smooth twin support vector machines via unconstrained convex minimization[J]. Filomat, 2017, 31(8): 2195-2210. DOI: 10.2298/fil1708195t .
[13]
YANH, YEQ L, ZHANGT A, et al. Least squares twin bounded support vector machines based on L1-norm distance metric for classification[J]. Pattern Recognition, 2018, 74(C): 434-447. DOI: 10.1016/j.patcog.2017.09.035 .
[14]
XIAOY S, LIUJ N, WENK R, et al. A least squares twin support vector machine method with uncertain data[J]. Applied Intelligence, 2023, 53(9): 10668-10684. DOI: 10.1007/s10489-022-03897-3 .
[15]
GUPTAU, GUPTAD. Bipolar fuzzy based least squares twin bounded support vector machine[J]. Fuzzy Sets and Systems, 2022, 449: 120-161. DOI: 10.1016/j.fss.2022.06.009 .
ZHENGX H, ZHANGL, YANL L. CTSVM: A robust twin support vector machine with correntropy-induced loss function for binary classification problems[J]. Information Sciences, 2021, 559: 22-45. DOI: 10.1016/j.ins.2021.01.006 .
[18]
DAMMINSEDV, PANUPW, WANGKEEREER. Laplacian twin support vector machine with pinball loss for semi-supervised classification[J]. IEEE Access, 2023, 11: 31399-31416. DOI: 10.1109/ACCESS.2023.3262270 .
[19]
HAZARIKAB B, GUPTAD. Density weighted twin support vector machines for binary class imbalance learning[J]. Neural Processing Letters, 2022, 54(2): 1091-1130. DOI: 10.1007/s11063-021-10671-y .
[20]
CHEZ Y, LIUB, XIAOY S, et al. A new twin SVM method with dictionary learning[J]. Applied Intelligence, 2021, 51(10): 7245-7261. DOI: 10.1007/s10489-021-02273-x .
[21]
STEINWARTI. Consistency of support vector machines and other regularized kernel classifiers[J]. IEEE Transactions on Information Theory, 2005, 51(1): 128-142. DOI: 10.1109/TIT.2004.839514 .
[22]
VAPNIKV N, CHERVONENKISA J. The necessary and sufficient conditions for consistency of the method of empirical risk minimization[J]. Pattern Recognition and Image Analysis, 1991, 1(3): 283-305.
[23]
VAPNIKV N. Statistical Learning Theory[M]. New York: Wiley, 1998.
[24]
VAPNIKV N. The Nature of Statistical Learning Theory[M]. 2nd ed. New York: Springer, 2000. DOI: 10.1007/978-1-4757-3264-1_8 .
[25]
ROYA, CHAKRABORTYS. Support vector machine in structural reliability analysis: A review[J]. Reliability Engineering & System Safety, 2023, 233: 109126. DOI: 10.1016/j.ress.2023.109126 .
[26]
XUW X, HUANGD J, ZHOUS G. Universal consistency of twin support vector machines[J]. International Journal of Machine Learning and Cybernetics, 2021, 12(7): 1867-1877. DOI: 10.1007/s13042-021-01281-0 .
[27]
LINS B, WANGK D, WANGY, et al. Universal consistency of deep convolutional neural networks[J]. IEEE Transactions on Information Theory, 2022, 68(7): 4610-4617. DOI: 10.1109/TIT.2022.3151753 .
[28]
DECOSTED, BURLM C. Distortion-invariant recognition via jittered queries[C]//Proceedings IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE Press, 2002: 732-737. DOI: 10.1109/CVPR.2000.855893 .
[29]
BOUSQUETO, ELISSEEFFA. Algorithmic stability and generalization performance[C]//Proceedings of the 13th International Conference on Neural Information Processing Systems. New York: ACM, 2000: 196-202. DOI: 10.5555/3008751.3008778 .