Multiclass Support Vector Machines Using Balanced Dichotomization
- 1.1k Downloads
The Support Vector Machine (SVM) has been introduced as a technique for solving a variety of learning and function estimation problems. The technique was originally designed for binary classification learning with its outstanding performance. However, many real world applications involve multiclass classification. Typical SVM solutions to N-class problems are to construct and combine several two-class classifiers into an N-class classifier such as the one-against-the-rest approach (1-v-r) and the one-against-one approach (1- v-1). The one-against-one methods solve N(N?1)/2 binary classifiers where each one is trained on data from two classes.
- 1.Bartlett, P.L., Shawe-Taylor, J.: Generalization performance of support vector machines and other pattern classifiers. In: Schölkopf, B., Burges, C., Smola, A. (eds.) Advances in Kernel Methods – Support Vector Learning, pp. 43–54. MIT Press, USA (1999)Google Scholar
- 2.Blake, C., Keogh, E., Merz, C.: UCI Repository of Machine Learning Databases, Department of Information and Computer Science, University of California, Irvine (1998), http://www.ics.uci.edu/~mlearn/MLSummary.html
- 3.Friedman, J.H.: Another Approach to Polychotomous classification, Technical report, Department of Statistics, Stanford University (1996)Google Scholar
- 4.Kijsirikul, B., Ussivakul, N., Meknavin, S.: Adaptive Directed Acyclic Graphs for Multiclass Classification. In: The Seventh Pacific Rim International Conference on Artificial Intelligence (2002)Google Scholar
- 5.Platt, J., Cristianini, N., Shawe-Taylor, J.: Large Margin DAGs for Multiclass Classification. In: Advances in Neural Information Processing Systems, vol. 12, pp. 547–553. MIT Press, Cambridge (2000)Google Scholar