Abstract
In this paper, while considering rejection as an option, we attempt to tackle the problem of multiclass classification and the uncertainty that arises from the class possibility assignment of data. To address the challenge of classification based on possible class assignments, we use the likelihood ratio, which helps us develop a holistic approach that considers all the positive and negative effects of assigning a particular class as opposed to others to a data point. To this end, we propose a possibilistic variant of the contrastive-learning function, inspired by RSLVQ [20], and a class-wise decision rule based on it. The latter is used to define the total cost function. In addition, with the help of likelihood ratio, an error-rejection trade-off inspired by Chow [3], is proposed. Finally, modification of the cost function and integration of rejection into it result in an interpretable model whose capabilities in both aspects (classification/rejection) are demonstrated by application to different data sets.
Supported by Die Bundesanstalt für Arbeitsschutz und Arbeitsmedizin (BAuA).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bousquet, O., Elisseeff, A.: Stability and generalization. J. Mach. Learn. Res. 2, 499–526 (2002)
Bradley, A.P.: The use of the area under the roc curve in the evaluation of machine learning algorithms. Pattern Recogn. 30(7), 1145–1159 (1997)
Chow, C.: On optimum recognition error and reject tradeoff. IEEE Trans. Inf. Theory 16(1), 41–46 (1970)
Devarakota, P.R., Mirbach, B., Ottersten, B.: Confidence estimation in classification decision: a method for detecting unseen patterns. In: Advances in Pattern Recognition, pp. 290–294. World Scientific (2007)
Fischer, L., Nebel, D., Villmann, T., Hammer, B., Wersing, H.: Rejection strategies for learning vector quantization – a comparison of probabilistic and deterministic approaches. In: Villmann, T., Schleif, F.-M., Kaden, M., Lange, M. (eds.) Advances in Self-Organizing Maps and Learning Vector Quantization. AISC, vol. 295, pp. 109–118. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-07695-9_10
Hammer, B., Villmann, T.: Generalized relevance learning vector quantization. Neural Netw. 15(8–9), 1059–1068 (2002)
Kohavi, R., et al.: A study of cross-validation and bootstrap for accuracy estimation and model selection. In: IJCAI, Montreal, Canada, vol. 14, pp. 1137–1145 (1995)
Kohonen, T.: Learning vector quantization. In: Kohonen, T. (ed.) Self-Organizing Maps, pp. 175–189. Springer, Heidelberg (1995). https://doi.org/10.1007/978-3-642-97610-0_6
Kohonen, T., Hynninen, J., Kangas, J., Laaksonen, J., Torkkola, K.: LVQ PAK: the learning vector quantization program package. Technical report (1996)
Kullback, S., Leibler, R.A.: On information and sufficiency. Ann. Math. Stat. 22(1), 79–86 (1951)
Musavishavazi, S., Mohannazadeh Bakhtiari, M., Villmann, T.: A mathematical model for optimum error-reject trade-off for learning of secure classification models in the presence of label noise during training. In: Rutkowski, L., Scherer, R., Korytkowski, M., Pedrycz, W., Tadeusiewicz, R., Zurada, J.M. (eds.) ICAISC 2020. LNCS (LNAI), vol. 12415, pp. 547–554. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-61401-0_51
Musavishavazi, S., Kaden, M., Villmann, T.: Possibilistic classification learning based on contrastive loss in learning vector quantizer networks. In: Rutkowski, L., Scherer, R., Korytkowski, M., Pedrycz, W., Tadeusiewicz, R., Zurada, J.M. (eds.) ICAISC 2021. LNCS (LNAI), vol. 12854, pp. 156–167. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-87986-0_14
Nebel, D., Hammer, B., Villmann, T.: A median variant of generalized learning vector quantization. In: Lee, M., Hirose, A., Hou, Z.-G., Kil, R.M. (eds.) ICONIP 2013. LNCS, vol. 8227, pp. 19–26. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-42042-9_3
Provost, F., Fawcett, T.: Robust classification for imprecise environments. Mach. Learn. 42(3), 203–231 (2001)
Raschka, S.: Model evaluation, model selection, and algorithm selection in machine learning. arXiv preprint arXiv:1811.12808 (2018)
Ravichandran, J., Kaden, M., Saralajew, S., Villmann, T.: Variants of dropconnect in learning vector quantization networks for evaluation of classification stability. Neurocomputing 403, 121–132 (2020)
Robbins, H., Monro, S.: A stochastic approximation method. Ann. Math. Stat. 400–407 (1951)
Saralajew, S., Holdijk, L., Rees, M., Asan, E., Villmann, T.: Classification-by-components: probabilistic modeling of reasoning over a set of components. In: Advances in Neural Information Processing Systems, vol. 32 (2019)
Sato, A., Yamada, K.: Generalized learning vector quantization. In: NIPS, vol. 95, pp. 423–429 (1995)
Seo, S., Obermayer, K.: Soft learning vector quantization. Neural Comput. 15(7), 1589–1604 (2003)
Varma, S., Simon, R.: Bias in error estimation when using cross-validation for model selection. BMC Bioinform. 7(1), 1–8 (2006)
Villmann, A., Kaden, M., Saralajew, S., Hermann, W., Villmann, T.: Reliable patient classification in case of uncertain class labels using a cross-entropy approach. In: ESANN (2018)
Villmann, A., Kaden, M., Saralajew, S., Villmann, T.: Probabilistic learning vector quantization with cross-entropy for probabilistic class assignments in classification learning. In: Rutkowski, L., Scherer, R., Korytkowski, M., Pedrycz, W., Tadeusiewicz, R., Zurada, J.M. (eds.) ICAISC 2018. LNCS (LNAI), vol. 10841, pp. 724–735. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-91253-0_67
Villmann, T., et al.: Self-adjusting reject options in prototype based classification. In: Merényi, E., Mendenhall, M.J., O’Driscoll, P. (eds.) Advances in Self-Organizing Maps and Learning Vector Quantization. AISC, vol. 428, pp. 269–279. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-28518-4_24
Villmann, T., Kaden, M., Nebel, D., Biehl, M.: Learning vector quantization with adaptive cost-based outlier-rejection. In: Azzopardi, G., Petkov, N. (eds.) CAIP 2015. LNCS, vol. 9257, pp. 772–782. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-23117-4_66
Zadeh, L.A.: Probability measures of fuzzy events. J. Math. Anal. Appl. 23(2), 421–427 (1968)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Musavishavazi, S., Alipour, M. (2022). Possibilistic Reject-Classification Based on Contrastive Learning in Vector Quantization Networks. In: Tan, Y., Shi, Y. (eds) Data Mining and Big Data. DMBD 2022. Communications in Computer and Information Science, vol 1744. Springer, Singapore. https://doi.org/10.1007/978-981-19-9297-1_25
Download citation
DOI: https://doi.org/10.1007/978-981-19-9297-1_25
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-19-9296-4
Online ISBN: 978-981-19-9297-1
eBook Packages: Computer ScienceComputer Science (R0)