Abstract
Extending linear classifiers from feature vectors to attributed graphs results in sublinear classifiers. In contrast to linear models, the classification performance of sublinear models depends on our choice as to which class we label as positive and which as negative. We prove that the expected classification accuracy of sublinear models may differ for different class labelings. Experiments confirm this finding for empirical classification accuracies on small samples. These results give rise to flip-flop sublinear classifiers that consider both class labelings during training and select the model for prediction that better fits the training data.
Keywords
Download to read the full chapter text
Chapter PDF
Similar content being viewed by others
References
Bunke, H., Riesen, K.: Improving vector space embedding of graphs through feature selection algorithms. Pattern Recognition 44(9), 1928–1940 (2011)
Cour, T., Srinivasan, P., Shi, J.: Balanced graph matching. In: NIPS (2006)
Gold, S., Rangarajan, A.: A graduated assignment algorithm for graph matching. IEEE Trans. on Pattern Analysis and Machine Intelligence 18(4), 377–388 (1996)
Hastie, T., Tibshirani, R., Friedman, J.: The elements of statistical learning. Springer, New York (2001)
Jain, B., Wysotzki, F.: Multi-Layer Perceptron Learning in the Domain of Graphs. IJCNN 3, 1993–1998 (2003)
Jain, B., Wysotzki, F.: Structural perceptrons for attributed graphs. In: Fred, A., Caelli, T.M., Duin, R.P.W., Campilho, A.C., de Ridder, D. (eds.) SSPR&SPR 2004. LNCS, vol. 3138, pp. 85–94. Springer, Heidelberg (2004)
Jain, B.: Structural Neural Learning Machines. PhD thesis, TU Berlin (2005)
Jain, B., Obermayer, K.: Structure Spaces. The Journal of Machine Learning Research 10, 2667–2714 (2009)
Jain, B., Obermayer, K.: Maximum likelihood for gaussians on graphs. In: Jiang, X., Ferrer, M., Torsello, A. (eds.) GbRPR 2011. LNCS, vol. 6658, pp. 62–71. Springer, Heidelberg (2011)
Jain, B., Obermayer, K.: Generalized learning graph quantization. In: Jiang, X., Ferrer, M., Torsello, A. (eds.) GbRPR 2011. LNCS, vol. 6658, pp. 122–131. Springer, Heidelberg (2011)
Jain, B., Obermayer, K.: Learning in Riemannian Orbifolds. arXiv preprint arXiv:1204.4294 (2012)
Jain, B.: Sublinear Models for Graphs. arXiv preprint arXiv:1204.4294 (2014)
Jain, B.: Flip-Flop Sublinear Models for Graphs: Proof of Theorem 1. arXiv preprint arXiv:cs.LG (2014)
Livi, L., Rizzi, A., Sadeghian, A.: Optimized Dissimilarity Space Embedding for Labeled Graphs. Information Sciences (2014)
Riesen, K., Neuhaus, M., Bunke, H.: Graph embedding in vector spaces by means of prototype selection. In: Escolano, F., Vento, M. (eds.) GbRPR. LNCS, vol. 4538, pp. 383–393. Springer, Heidelberg (2007)
Riesen, K., Bunke, H.: IAM graph database repository for graph based pattern recognition and machine learning. In: da Vitoria Lobo, N., Kasparis, T., Roli, F., Kwok, J.T., Georgiopoulos, M., Anagnostopoulos, G.C., Loog, M. (eds.) SSPR&SPR 2008. LNCS, vol. 5342, pp. 287–297. Springer, Heidelberg (2008)
Umeyama, S.: An eigendecomposition approach to weighted graph matching problems. IEEE Transactions on PAMI 10(5), 695–703 (1988)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jain, B. (2014). Flip-Flop Sublinear Models for Graphs. In: Fränti, P., Brown, G., Loog, M., Escolano, F., Pelillo, M. (eds) Structural, Syntactic, and Statistical Pattern Recognition. S+SSPR 2014. Lecture Notes in Computer Science, vol 8621. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-44415-3_10
Download citation
DOI: https://doi.org/10.1007/978-3-662-44415-3_10
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-662-44414-6
Online ISBN: 978-3-662-44415-3
eBook Packages: Computer ScienceComputer Science (R0)