Abstract
We try to overcome the imbalanced data set problem in image annotation by choosing a convenient loss function for learning the classifier. Instead of training a standard SVM, we use a Ranking SVM in which the chosen loss function is helpful in the case of imbalanced data. We compare the Ranking SVM to a classical SVM with different visual features. We observe that Ranking SVM always improves the prediction quality, and can perform up to 23% better than the classical SVM.
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
Glotin, H., Fakeri-Tabrizi, A., Mulhem, P., Ferecatu, M., Zhao, Z.-Q., Tollari, S., Quenot, G., Sahbi, H., Dumont, E., Gallinari, P.: Comparison of various AVEIR visual concept detectors with an index of carefulness. In: CLEF Working Notes (2009)
Joachims, T.: A support vector method for multivariate performance measures. In: International Conference on Machine Learning, ICML (2005)
Nowak, S., Dunker, P.: Overview of the CLEF 2009 large scale — visual concept detection and annotation task. In: Peters, C., et al. (eds.) CLEF 2009 Workshop, Part II. LNCS, vol. 6242, pp. 94–109. Springer, Heidelberg (2010)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Fakeri-Tabrizi, A., Tollari, S., Usunier, N., Gallinari, P. (2010). Improving Image Annotation in Imbalanced Classification Problems with Ranking SVM. In: Peters, C., et al. Multilingual Information Access Evaluation II. Multimedia Experiments. CLEF 2009. Lecture Notes in Computer Science, vol 6242. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-15751-6_37
Download citation
DOI: https://doi.org/10.1007/978-3-642-15751-6_37
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-15750-9
Online ISBN: 978-3-642-15751-6
eBook Packages: Computer ScienceComputer Science (R0)