A Fast Approach to Improve Classification Performance of ECOC Classification Systems
Error correcting output coding is a well known technique to decompose a multi-class classification problem into a group of two-class problems which can be faced by using a combination of binary classifiers. Each of them is trained on a different dichotomy of the classes. The way the set of classes is mapped on this set of dichotomies may essentially influence the obtained performance. In this paper we present a new tool, the k-NN lookup table to optimize this mapping in a fast way and a fast procedure to change the dichotomies in a proper way. Experiments on artificial and public data sets show that the proposed procedure may significantly improve the ECOC performance in multi-class problems.
Keywordsmultiple classifiers systems ECOC k-NN
- 2.Kong, E.B., Dietterich, T.G.: Error-Correcting Output Coding Corrects Bias and Variance. In: International Conference on Machine Learning, pp. 313–321 (1995)Google Scholar
- 3.Kong, E.B., Dietterich, T.G.: Probability estimation via error-correcting output coding. In: Int. Conf. of Articial Inteligence and soft computing, Ban, Canada (1997)Google Scholar
- 4.Allwein, E.L., Schapire, R.E., Singer, Y.: Reducing multiclass to binary: A unifying approach for margin classifiers. In: Proceedings of the Seventeenth International Conference on Machine Learning, pp. 9–16 (2000)Google Scholar
- 5.Crammer, K., Singer, Y.: On the Learnability and Design of Output Codes for Multiclass Problems. Machine Learning 47 (2002)Google Scholar
- 9.Blake, C., Keogh, E., Merz, C.J.: UCI Repository of Machine Learning Databases (1998), http://www.ics.uci.edu/~mlearn/MLRepository.html