Abstract
We address the issue of fitting a logistic regression model to soft labeled data, when the soft labels take the form of plausibility degrees for the classes. We propose to use the E2M algorithm to take this partial information into account. The resulting procedure iterates two steps: first, expected class memberships are computed using the soft labels and the current parameter estimates; then, new parameter estimates are obtained using these expected memberships. Experimental results show the interest of our approach when the data labels are corrupted with noise.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Ambroise, C., Govaert, G.: EM algorithm for partially known labels. In: 7th Conference of the International Federation of Classification Societies (IFCS 2000), pp. 161–166 (2000)
Amini, M.R., Gallinari, P.: Semi-supervised logistic regression. In: 15th European Conference on Artificial Intelligence (ECAI 2002), pp. 390–394 (2002)
Bishop, C.M.: Pattern Recognition and Machine Learning. Springer (2006)
Chapelle, O., Schšlkopf, B., Zien, A.: Semi-supervised learning. The MIT Press, Cambridge (2006)
Cour, T., Sapp, B., Taskar, B.: Learning from Partial Labels. Journal of Machine Learning Research 12, 1501–1536 (2011)
Denœux, T.: Maximum likelihood estimation from Uncertain Data in the Belief Function Framework. IEEE Transactions on Knowledge and Data Engineering 25(1), 119–130 (2013)
Erkan, A.N., Altun, Y.: Semi-Supervised Learning via Generalized Maximum Entropy. In: 13th International Conference on Artificial Intelligence and Statistics (AISTATS 2010), pp. 209–216 (2010)
Grandvalet, Y.: Logistic regression for partial labels. In: 9th International Conference on Information Processing and Management of Uncertainty in Knowlege-based Systems (IPMU 2002), vol. III, pp. 1935–1941 (2002)
Grandvalet, Y., Bengio, Y.: Semi-supervised learning by entropy minimization. In: Neural Information Processing Systems 17 (NIPS 2004), pp. 529–536. The MIT Press (2005)
Hüllermeier, E., Beringer, J.: Learning from ambiguously labeled examples. In: Famili, A.F., Kok, J.N., Peña, J.M., Siebes, A., Feelders, A. (eds.) IDA 2005. LNCS, vol. 3646, pp. 168–179. Springer, Heidelberg (2005)
Hüllermeier, E.: Learning from imprecise and fuzzy observations: Data disambiguation through generalized loss minimization. International Journal of Approximate Reasoning, DOI: 10.1016/j.ijar.2013.09.003 (2013) (in Press)
Li, J., Bioucas-Dias, J.: Semisupervised Hyperspectral Image Classification using Soft Sparse Multinomial Logistic Regression. IEEE Geoscience and Remote Sensing Letters, 318–322 (2013)
Nguyen, Q., Valizadegan, H., Hauskrecht, M.: Learning classification models with soft-label information. J. Am. Med. Inform. Assoc. (2014), doi:10.1136/amiajnl-2013-001964
Shafer, G.: A mathematical theory of evidence. Princeton University Press (1976)
Smets, P., Kennes, R.: The Transferable Belief Model. Artificial Intelligence 66, 191–243 (1994)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Quost, B. (2014). Logistic Regression of Soft Labeled Instances via the Evidential EM Algorithm. In: Cuzzolin, F. (eds) Belief Functions: Theory and Applications. BELIEF 2014. Lecture Notes in Computer Science(), vol 8764. Springer, Cham. https://doi.org/10.1007/978-3-319-11191-9_9
Download citation
DOI: https://doi.org/10.1007/978-3-319-11191-9_9
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11190-2
Online ISBN: 978-3-319-11191-9
eBook Packages: Computer ScienceComputer Science (R0)