Efficient Pairwise Classification Using Local Cross Off Strategy
The pairwise classification approach tends to perform better than other well-known approaches when dealing with multiclass classification problems. In the pairwise approach, however, the nuisance votes of many irrelevant classifiers may result in a wrong prediction class. To overcome this problem, a novel method, Local Crossing Off (LCO), is presented and evaluated in this paper. The proposed LCO system takes advantage of nearest neighbor classification algorithm because of its simplicity and speed, as well as the strength of other two powerful binary classifiers to discriminate between two classes. This paper provides a set of experimental results on 20 datasets using two base learners: Neural Networks and Support Vector Machines. The results show that the proposed technique not only achieves better classification accuracy, but also is computationally more efficient for tackling classification problems which have a relatively large number of target classes.
KeywordsMulticlass Pairwise classification Neural Networks Support Vector Machines
Unable to display preview. Download preview PDF.
- 3.Blake, C., Merz, C.: Uci repository of machine learning databases, department of information and computer sciences, university of california, irvine (1998)Google Scholar
- 4.Chang, C.C., Lin, C.J.: Libsvm: A library for support vector machines (2001)Google Scholar
- 10.Fei, B., Liu, J.: Binary tree of svm: A new fast multiclass training and cassification algorithm. IEEE Transactions on Neural Networks 17(696-704) (2006)Google Scholar
- 14.Hastie, T., Tibshirani, R.: Classification by pairwise coupling. In: Proceedings of the 1997 Conference on Advances in Neural Information Processing Systems 10, NIPS 1997, pp. 507–513. MIT Press, Cambridge (1998)Google Scholar
- 17.Iman, R., Davenport, J.: Approximations of the critical regions of the friedman statistic. Communications in Statistics 6, 571–595 (1980)Google Scholar
- 18.Knerr, S., Personnaz, L., Dreyfus, G.: Single-layer learning revisited: a stepwise procedure for building and training a neural network. In: Fogelman, J. (ed.) Neurocomputing: Algorithms, Architectures and Applications. Springer (1990)Google Scholar
- 23.Platt, J.C., Cristianini, N., Shawe-taylor, J.: Large margin dags for multiclass classification. Advances in Neural Information Processing Systems 12, 547–553 (2000)Google Scholar